Jchess Tutorial 3.0

User’s ManualLHM/RD/02/13

CHESS Tutorial and CookbookUpdated for version 3.0

Ecole des Mines de ParisCentre d’Informatique Geologique

Fontainebleau, France

The user’s guide for the geochemical model CHESS and its Graphic User Interface, JCHESS. Withpractical examples and useful tips to help scientists and engineers in solving common problems inhydro-geochemistry and environmental sciences in general.

April 2002

Ecole Nationale Superieure des Mines de Paris Report Nr. LHM/RD/02/13Centre d’Informatique Geologique

CHESS Tutorial and CookbookUpdated for version 3.0

Jan van der LeeLaurent De Windt

April 2002

CHESS/JCHESS Copyrightc© 1993-2002 ARMINES

Centre d’Informatique Geologique of theEcole des Mines de Paris35 rue Saint-Honore, 77305 Fontainebleau, France.All rights reserved.

Reference type :J. van der Lee and L. De Windt (2002). CHESS Tutorial and Cookbook. Updated for version 3.0.UsersManualNr. LHM/RD/02/13, Ecole des Mines de Paris, Fontainebleau, France.

CHESS stands for CHemical Equilibrium Speciation with SurfacesJCHESS, the graphical user interface of CHESS, stands for Java CHESSAll software and design by Jan van der Lee

CHESS is on the internet :http://chess.ensmp.frQuestions, bug reports or comments :[email protected] CHESS E-Mail list :chess [email protected]

First print for JCHESS version 2.0, CHESS version 3.0.

This document was prepared with free software : Linux/KDE, LATEX, Gimp.

Table des matieres

1 Introduction 1

1.1 Geochemistry, CHESS and JCHESS . . . . . . . . . . . . . . . . . . . . .. . . 1

1.2 Essential notions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 4

1.3 About this guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 5

2 A walk through Speciation Mode 7

2.1 Main solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

2.2 Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.3 Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.4 Sorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

2.5 Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

2.6 Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

2.7 Piper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.8 JPlot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

3 A walk through Diagram Mode 36

3.1 Water reduction and oxidizing . . . . . . . . . . . . . . . . . . . . . .. . . . . 37

3.2 Pourbaix diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 38

3.3 Adding more generality . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 41

3.4 Improvement of graphical aspects of the diagram . . . . . . .. . . . . . . . . . 41

4 Getting started 45

4.1 Running your first example . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 45

4.2 Quartz in pure water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 46

5 Carrying on 49

5.1 Solubility of ferric iron . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 49

5.2 Precipitation and dissolution . . . . . . . . . . . . . . . . . . . . .. . . . . . . 52

5.3 Redox disequilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 54

6 Sorption and ion exchange 57

6.1 Surface complexation . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 57

i

Table des matieres

6.2 Cation exchange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 60

6.3 TheKd approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

7 Reaction paths 67

7.1 Titration and speciation diagrams . . . . . . . . . . . . . . . . . .. . . . . . . . 67

7.2 Polythermal reaction paths . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 76

7.3 Flushing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76

7.4 Mixing of two solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 81

8 Kinetics of dissolution and precipitation 83

8.1 Kinetic laws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83

8.2 Calcite dissolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 84

8.3 Plagioclase albitization . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . 85

8.4 Temperature dependence . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 89

8.5 Catalyzing and inhibiting effects . . . . . . . . . . . . . . . . . .. . . . . . . . 90

A Thermodynamic databases 93

A.1 Available databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 93

A.2 The CHESS database format . . . . . . . . . . . . . . . . . . . . . . . . . .. . 96

A.3 Comments and references . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . 101

Bibliography 103

ii

Chapitre 1. Introduction

1.1 Geochemistry, CHESS and JCHESS

Aquatic geochemical systems form a complex ensemble of simultaneously interacting compo-nents. They contain elements as divers as chemical species in solution, mobile and immobile mi-neral phases, organic and inorganic components and gases. Geochemical models are used to shedlight on the behaviour of these elements and of aquatic systems in general. They are applied to awide range of problems in the laboratory (interpretation ofbatch- and column experiments) andin the field (water quality, contaminant fate, waste containment, acid mine drainage, hydrother-mal systems, ore deposits, petroleum reservoirs, etc.). They are used to address environmentalproblems such as the solubility of contaminants, their reactivity with respect to a soil or rockmatrix and their ability to complex with other aqueous species. But aquatic chemistry modellingalso helps in understanding other, less rocky (in geological sense) systems such as the humanbody : the chemistry of blood (radiotoxology), drinking water purification plants and electricpower stations (chemical corrosion) are just a few examples.

From a more formal view point, geochemical models are specialized calculation tools whichsolve large and highly non-linear, multi-dimensional setsof algebraic equations to calculate theequilibrium state or a kinetic reaction of the aquatic system. They also read, process and unders-tand extended databases which contain thermodynamic properties of the species. They providea report resuming the system in terms of hydro-geochemical quantities, characteristic parame-ters and other, useful information which is needed by scientists and engineers. More sophistica-ted models also provide graphical tools as a help to the interpretation of the numerical results.CHESS, which stands forCHemicalEquilibrium withSpecies andSurfaces, is such a computermodel and deals with a large number of processes and features.

CHESS is designed to meet demands of the most complexgeochemical modelling needs, yet it also aims to meet re-quirements of modern user-friendliness. Therefore, and sinceversion 2.5, CHESS has been extended with JCHESS, a gra-phical interface written in java. The java programming lan-guage has the important advantage of being portable to mostplatforms currently used : Windows, MacOS, Linux, Solaris...The graphic user interface JCHESS not only greatly impro-ved the user-friendliness of the model, it also increases theproductivity of the modelling scientist by displaying reactionpaths graphically and creating clear reports. JCHESS brings geochemistry to your desktop andfingertips : it is the first geochemical model which does not require the awkward learning processof many keywords which has to be typed according to a specific,rigid syntax. JCHESS allows fora quick start, to navigate between different databases, to modify species and to enrich a databasewith new species.

The current version of JCHESS proposes two distinct services : thespeciation modeand the

1


diagram mode. In speciation mode, some of the processes which can be simulated are :

• thermodynamic equilibrium with mineral, gaseous or colloidal phasesCHESS calculates the equilibrium state of the solution using the total concentrations orspecific activities to constrain the system. For example, itis straightforward to calculatethe equilibrium state with respect to one or several minerals, and/or with an imposed pHor pCO2.

• oxidation and reductionFor a given redox-potential, the model calculates the equilibrium state of the solution withrespect to the correct oxidation states. Redox may be triggered in different ways, e.g. im-posing the oxygen fugacity or providing the Eh or pe of the solution. In case of redoxdisequilibrium, it is possible to decouple specific speciesfrom the redox reactions. Also,CHESS allows to specify a redox-couple which sets the overall redox potential.

• precipitation and dissolution of mineral phasesCHESS precipitates minerals when the solution becomes over-saturated and recalculatesthe equilibrium state, this time with respect to the newly formed solid phases. Dissolutionwill occur when the mineral is undersaturated. This processis repeated until none of thephases is oversaturated. This feature can be turned off by the user, or fine-tuned usingkinetic reactions.

• formation and dissolution of colloidal phasesColloids may be formed if they appear to be oversaturated with respect to the aqueouscomposition. When undersaturated, colloids will dissolve. This feature can be turned on oroff by the user, or fine-tuned using kinetic reactions.

• surface complexationCalculation of interaction between aqueous species and reactive surface groups of mineralor colloidal species present in solution is based on the surface complexation theory withelectrostatic correction. Current electrostatic models are the constant capacitance modeland double- or triple layer models.

• cation exchangeCation exchange with minerals or colloids is possible instead of (or together with) surfacecomplexation. For example, clay particles may contain cation exchange and surface com-plexation sites and CHESS will calculate the equilibrium state taking into account bothtypes of site.

• multi-site reactionsA colloidal (organic or inorganic) or mineral surface may contain more than one type ofsite. As a matter of fact, the number of types of site per surface is unlimited. Thecontinuoussite distribution modelcan be emulated using a large number of sites.

• multi-surface reactionsThe number of reactive mineral- or colloid surfaces in the system is unlimited, which is auseful feature for complex, reactive geochemical systems where several surfaces competefor metals.

• reaction pathsReaction paths may be specified to bring the solution from onestate to another. Processes

2

1.1. Geochemistry, CHESS and JCHESS

such as titration, flushing or mixing of two solutions are available and easy to use. Theresults of reaction path calculations are graphically displayed.

• temperature dependence (0 to 300C)The solution temperature can be set to any value above 0C. The database contains poly-nomial values up to 300C for most species.

• kinetic control of dissolution and precipitationKinetic control of dissolution and precipitation of mineral and colloidal phases can beenabled for any period of time through a general law including the reactive surface area,catalyzing or inhibiting effects and temperature dependence.

• Piper diagramsPiper diagrams are often used to graphically qualify a solution in terms of major cationsand anions. CHESS allows to display a Piper diagram of one or several solutions and evenof an entire reaction path.

These functions —there are many more— will be outlined in detail in the following chapters,illustrated with practical examples.

In diagram mode, the user can create different kind of activity-diagrams, i.e. a graph from whichone can read the expected predominant species as a function of two variables. Diagram typesproposed are :

• Pourbaix diagramsAs a function of pH and Eh, pe or oxygen fugacity, the predominant species of a specifiedelement

• activity-activity diagramsDiagrams showing the predominant species of a specified element as a function of anarbitrary choice of species, i.e. HCO3

−, SO42−,. . ..

• mosaic diagramsOne may add one or more species to the system, which will be speciated according to theaxes variables. Mosaic diagrams are much more precise than classical Pourbaix diagramsif auxiliary species are dependent on one of the axes variables.

• solubility diagrmsSpecifying a diagram species as a function of its activity onthe y-axis and, e.g., pH or Eh,on the x-axis leads to a typical solubility diagram. Solubility are often used to indicate thesolubility-determining phases of a metal as a function of pH.

There are a multitude of graphical and mathematical possibilities in diagram mode. Severalexamples will be illustrate these in one of following chapters.

As follows from this list, CHESS fulfills the general modelling-desiderata. More specifically, itadds (colloidal and mineral) surface interface processes,options to handle heterogeneous reactivesystems, reaction path modeling and tools to graphically display simulation results, piper andactivity diagrams. CHESS comes with several well known databases, all rewritten in a clearlystructured, flexible format. The default database,chess.tdb, is based on the full LLNL databasebut extended with colloids, organic matter and complexation sites. Appendix A of this manualdescribes the different available databases and explains the common database format in detail.

3


CHESS is written in C++ and is entirely object-oriented, which improves the readability, mainte-nance and development of the code. It is important to know that CHESS has been developed forcoupling purposes with hydrodynamic models. In view of thisgoal, all basic functions relatedto database reading and treatment as well as the Newton-Raphson solver are grouped together ina library, CHESSLIB. This library can be linked to any hydrodynamic model and provides ac-cess to equilibrium concentrations, sorbed or precipitated quantities and other useful information.Currently, CHESS is coupled to several hydrodynamic codes where among HYTEC (van der Leeet al. 2002), CHEMTRAP (Lucille et al. 2000) and hydro-mechanical code of the French CEA,CASTEM.

Notwithstanding the nice envelop, CHESS remains and will always remain specialist software forwhich some basic knowledge is required. Before we enter the domain of modelling nitty-grittyand geochemistry, let’s give a short overview of how JCHESS actually works.

1.2 Essential notions

Computers calculate faster than humans, but they wouldn’t pass the simplest test on communica-tion skills : a computer will, under the best circumstances,silently ignore a question like ”I wouldlike to know how much of 50 grams of kaolinite dissolves over 10 years please”. Maybe this willchange in the near future, but meanwhile, we have to adapt ourhuman language to the compu-ter. Fortunately, CHESS quite well understands something like a sequence of commands, i.e.,mineral Kaolinite = 50 g, time = 10 yr : less prosaic but reasonably close to commonEnglish.

Hence, the user has to learn the meaning of a minimum number ofessential quantities suchasconcentration, total-concentration andaqueous-concentration : qualifiers used tospecify what the quantity introduced actually represents.Similarly, in order to fully exploit themodel’s capacities, the user should be aware of the differences between e.g. titration, mixing andflushing, understanding the meaning of an imposedfugacity, and at least some of the options.And last but not least, the modelling scientist should be able to extract the required informationfrom the output files (orreport, in CHESS parlance). We explain these notions by means of atour through the JCHESS panels. Each panel, each option and particularity of CHESS will beoutlined. Specific examples, however, are to be found in following chapters which approach themodel from a purely geochemical view point.

JCHESS displays main panel with different sub-panels, menuitems and a toolbar. The menuitems are, in order,file, models, actions, settings andabout. Thefile menu allows toload a project script-file, tosavethe current project, toprint the script or toquit. When openinga project, JCHESS looks, by default, for files with the.chsextension. It has become a traditionamong CHESS users to use.chsas the extension of CHESS input scripts, i.e., something likeinput.chs. We recommend to continue this tradition.

The second menu item,models, allows to select the major model mode. Currently availableare :

• Speciation mode (default)This mode is what you classically expect from a speciation model : it allows you to set up

4

1.3. About this guide

a geochemical system, to calculate the equilibrium state, to perform reaction paths, displaythe results graphically and compare with analytical data, etc..

• Diagram modeThis mode allows to build activity-activity diagrams such as Pourbaix, solubility and mo-saic diagrams1.

Both modelling modes are fundamentally different : if you have made unsaved changes the modelwill ask to save the data before switching to another mode, since all changes will be lost.

The menu itemactions contains several specific actions of the code, such asrun (run chess,either for a speciation calculation or a diagram build, depending on the current mode) andshowreport, displaying the report with the equilibrium results in the JCHESS viewer. After a run,CHESS automatically updates graphs and reports currently opened on the desktop.

The menu itemsettings allows to change some default settings of the global modelling envi-ronment : the CHESS executable actually used (don’t change the default unless you know exactlywhat you are doing !) and numerical parameters, but also thelook & feel of the current appea-rance of JCHESS. There is also a submenu to display the CHESS console which comes in handyunder certain circumstances.

But what exactly is the console ? The console contains information about the actual modellingactions : it says how the database is parsed, informs about events during parsing and during equi-librium calculations, etc.. (see Figure 1.1). Normally, all this information is not very importantfor scientists, merely concerned by geochemical issues. Those who are interested, however, canopen the console after a run and have a look at it.

The console becomes, however, an essential and extremely useful help if CHESS exits anoma-lously : the database cannot be found, there is something wrong with the settings or CHESSencountered a problem to converge to the right solution. Even more important : the consoleprovides hints to solve the problem. To solve the problem, follow the advise following the errormessage, if provided. The console is quiet by default : a moreverbose output can be obtained viatheOutput panel (section 2.6). Note that under certain conditions, CHESS fails but the consoledoesn’t pop up. It is good practice to open the console manually (i.e. via thesettings menu) ifCHESS is behaving strangely (e.g. the graph is not displaying the expected results or the reportis not complete).

Some of the most currently used menu items are accessible viashortcut buttons which you canfind on the toolbar. The toolbar can be detached from the application, or placed at other sides ofthe main JCHESS window.

1.3 About this guide

The following two chapters will guide you through the two major modelling modes of CHESS.Progressively you will learn how to construct geochemical system, how to fine-tune the modelwith specific options and how to achieve the information you are looking for. Most menus are

1If you prefer to start JCHESS in diagram mode by default, launch the model with option-a (a = activity-diagrams).

5


FIG . 1.1– CHESS console. This window intercepts all error messages from CHESSand provides hints or remarks in case of a convergence failure.

illustrated with captures of the screen. Note that, depending on the operating system used, theactualLook & Feelsuch as window decoration, button styles etc. change. All figures in this guidecorrespond the Metal Look & Feel, the default on Linux platforms (which has been used to writethe software, as well as this guide).

The remaining chapters have been conceived to progressively introduce the reader to the worldof geochemical modelling with help of practical examples related to laboratory experiences orgeochemical systems. The more fundamental and theoreticalaspects are the subject of a comple-mentary document (van der Lee 1998).

6

Chapitre 2. A walk through Speciation Mode

Aqueous speciation, precipitation/dissolution processes, adsorption and reaction path simulationsare done in this mode. Simulating systems in speciation modeis what CHESS was originallydesigned for, and this chapter is an attempt to learn you how to set up a model of an aquaticsystem.Speciation mode is selected in themodels menu item : it is the default mode.

2.1 Main solution

Most of the time, you will be working with a single aquatic system or solution, the so-calledmain solution. The first panel of JCHESS (Figure 2.1) defines the main solution in terms of thefollowing items :

• A list of species (aqueous species, minerals, colloids, gases) :Species are added via the concentration editor which shows up when you click onadd....Double-clicking on a row allows you to modify the qualifier, concentration or unit. Anotherway of adding, deleting or editing species is to right-clickon either a row, either in thetable area to open a contextual pop-up-window. Species can be toggled on and off usingthe check-box in the left-most column. When toggled off, thespecies is not taken intoaccount by the calculations. The usefulness of the toggle option will become evident veryquickly, i.e. to test the impact of the presence of a certain species, without the need to addor delete it several times.

• Variables such as pH, Eh or pe :These are added as species (see above), hence via the concentration editor. They are pre-sented in a list of ’variables’.

• the temperature of the solution

• Volume :Essential when using units like ’gram’ or ’mol’ for species,or when simulating flushing ormixing reactions. By default, CHESS uses a total concentration of H2O of 1 kg, leading toa volume of approximately 1 liter.

• Time :If kinetics are involved, you must set the (initial) equilibration time of the solution. Withoutkinetically controlled reactions, the value defined in the main solution panel is ignored.

• Balance :it is possible to specify a species for which its concentration is adjusted by CHESS toobtain or maintain theelectroneutralityof the solution, called thebalance species. Bydefault, the balance feature is disabled.

• Density of the solution :You may either ask the model to evaluate the density (checkfree) or impose the densityof the solution.

7


• Redox state :In contrast to other geochemical models, redox is disabled by default. This is for historicalreasons, since CHESS was, in the first place, developed for modelling laboratory experi-ments where redox equilibrium was seldom reached. Notwithstanding this rule, redox istriggered automatically when either the Eh or pe of the solution is set. The redox must betriggered manually if you want to equilibrate the solution with, e.g., O2(g).

An interesting possibility of CHESS is to fix the redox potential via a redox-couple. Forexample, a common situation is a redox potential set by redoxcouple Fe3+/Fe2+. Whenusing a couple to fix the redox potential, make sure that both species composing the coupleare actually present in solution (Fe3+ and Fe2+, in this case).

Of course, it is possible and straight-forward to disregardthe alternative redox states of aspecies. In CHESS parlance, we speak aboutdisabling the redox for a species. Disablingthe redox of e.g. species NO3

− actuallydecouplesthis species from the redox reactions.

• Activity-correction models :Different models for non-ideality correction are proposedby CHESS, where among the

truncated Davies formulation (the ionic-strength dependency is truncated above an io-nic strength of 0.3), the unmodified Davies equation, the (extended) Debye-Huckel andthe B-dot formulæ. The truncated Davies formulation is the most commonly used model,while the B-dot is more precise for higher ionic strength andtemperatures other than 25C(van der Lee 1998).

One of the selectable items isnone, which disables calculation of activity corrections alto-gether. This is useful only in conjunction with a dedicated thermodynamic database, validfor a specific ionic strength.

CHESS allows to set the activity model used by the solvent, H2O, as well. Currently, onlythe Helgeson model is implemented. The solvent activity calculation is disabled by defaultsuch thatactH2O = 1 for all times.

Species or variables are added to the table via a specific purpose editor, theConcentrationeditor. Most editors of the JCHESS application behave and look similar to this editor : one ofthe reasons for which we will provide the following in-depthexplanation.

2.1.1 Concentration editor

Variables and species are selected using theConcentration editor. The editor pops-up whenyou click onadd... of the main solution panel, double-click on a row of the species table orright-click in the table and select eitheredit oradd (see Figure 2.2). This specific-purpose editorhas three major fields :

• Contextsall CHESS variables and species are arranged according to some kind of classification.

The rules used for classification are, to some extend, based on the database structure. Thefact is that these contexts help in finding species or items faster than building one long listof thousands of items. The editor distinguishes 7 differentcontexts :

8

2.1. Main solution

FIG . 2.1– Main solution, defining the system components and variables.

FIG . 2.2– Concentration editor, used to select species or variablesand to set theirvalue.

9


– variablessuch as pH, Eh and pe

– basis-speciesa list of all basis species, a subset of aqueous species in which you will generally findthe one you are looking for. See chapter A for the list of basisspecies of the CHESSdefault database.

– aqueous-speciesa list of all the aqueous species (including basis- and redoxspecies)

– colloidsinorganic colloidal species

– organicsorganic colloidal species (dissolvedorganic species are listed in the context ofaqueous-species,if the database contains them)

– gasesthe gases, for which you can set a fugacity

– mineralsall the minerals

• Itemsall the items available for the chosen context. There are only 3 items for thevariables

context, many more when you click on, e.g.,minerals. The exact number of items is afunction of the database used. For example, CHESS’ default databasechess.tdbcontainsabout 1220 aqueous species, 3 colloids, 3 organics, 91 gasesand 1126 minerals (the exactnumbers may change as we update the database on a regular basis).You can find a species in the list using the scrollbar (or the down-array or down-page key).A more efficient way of finding a species in a long list, however, is to usestart totype thefirst letters of the name of the species in theitem field of the selection (see below). Thelist will jump close to the item and you only have to complete the name by clicking on thefull name.

• SelectionThe right-most field allows to specify what type of quantity you would like to add, the

amount (with its unit) and, eventually, the surface area if the item is an inorganic colloid,an organic colloid or a mineral. Let us outline each field in some detail.Thequantitydepends on the actual context. Aqueous species (including basis- and redoxspecies) can be qualified by their

– total-concentrationcorresponding to the sum of the concentrations of all species containing the principalelement of the selected species1. For example, let’s consider element Ca representedby species Ca2+ :

1Although close to the facts, this statement is only approximatively true. For an exact and detailed explanation of themeaning oftotal-concentration, the reader is referred to theThermodynamic and mathematical concepts of CHESS(van der Lee 1998).

10

2.1. Main solution

total-concentration of Ca2+ = 1 mmolal

actually means

concentration of Ca2+ + concentration of CaOH+ +concentration of calcite + . . .= 1 mmolal

Then, the model calculates the actual concentrations (and activities) of each indivi-dual species.

– aqueous-concentrationcorresponding to the sum the concentrations of allaqueousspecies containing the

principal element of the selected species. Picking up the example used for the totalconcentration of element Ca, the aqueous-concentrationwould be defined as follows :

concentration of Ca2+ + concentration of CaOH+ + . . .= 1 mmolal

hence without calcite —which is not an aqueous species indeed.

– concentrationcorresponding to the actual concentration of the species, e.g. the concentration of

species Ca2+, or of species CaOH+. Consequently, the model will calculate the total-concentration of the principal element of the species.

– activityrarely used in speciation mode, this qualifier fixes the activity of the species and cal-culates the actual concentration and corresponding total concentration. The activityof a species is the product of the concentration and an activity coefficient, the lattercalculated with the specifiedactivity correction model.

Theminerals context proposes only two qualifiers :

– mineralindicating that the species is a mineral. This qualifier is basically identical to thetotal-concentration : it does not fix the actual concentration of the mineral, butthe (total) concentration of all species composing the mineral. Note that adding 1molal of a mineral does not necessarily give you 1 molal of thespecific mineral insolution ! Think about a spoon of sugar in your coffee. Part ofthe sugar will readilydissolve, leading to less than a spoon on the bottom of your cup. If you absolutelywant to have 1 molal of mineral, you should use specifierconcentration instead.

– concentrationcorresponding to the actual concentration of the mineral species, adjusting the totalconcentrations of the composing species, if necessary.

Similarly, the colloids and organics contexts propose two qualifiers,colloid andconcentration or organic andconcentration respectively. These qualifiers behaveexactly like for the mineral species. And finally, thegases context proposes two quali-fiers :

– fugacityfixes the fugacity of the gas on the atmosphere basis. For example, the fugacity of

CO2(g) in the atmosphere is∼3.16×10−4. Fugacity is dimensionless.

11


– total-concentrationTotal concentrations are fixed by aqueous components (basis-species, in general)hence this qualifier should not be used.

Thevalueof the quantity (activity, concentration) should be typed and a unit set. CHESSaccepts most commonly used units in geochemistry, such as molality (mol per kg H2O),molarity (mol per liter of solution), g/l, mg/l... CHESS also accepts mass-units such asmoles, kg, g... which are converted to molal units using the mass of the solvent, 1 kgby default. It is possible to set the mass of the solvent to another value by setting theconcentration of (basis) species H2O to e.g. 2 kg, or via the solution volume.

Thesurfacecan be set for all surface-bearing quantities, i.e., minerals and colloidal spe-cies. The specific surface of a species is used by kinetic lawsbut also to evaluate the reacti-vity of the surface in case of surface complexation reactions. There are two fundamentallydifferent ways to set the surface area :

– a specific surfaceproprement dit, in m2/kg : using this unit sets the surface to bea function of the actual mass of the solid. Precipitation or crystal-growth thereforeincreases the surface area (and therefore the precipitation rate and complexation ca-pacity) ;

– a volumetric surface area, in m2 per unit volume of solution : this unit makes thesurface independent of the solid concentration.

It is not always necessary to set the surface of a solid. If notset, the model uses defaultvalues. For mineral solids, CHESS assumes a single piece of mineral in spherical formwhich allows to calculate a minimum surface area. For inorganic and organic colloids,CHESS uses the default radius provided by the database to estimate a surface area.

Note : once you have finished adding the species to the solution, you must close the concentrationeditor. It has been decided this way in order to avoid confusion with other, sometimes closelyresembling editors. The editor is closed by a click on theDismiss button.

At this stage, you should be capable of building a solution with JCHESS. If you are too impatientto finish the JCHESS tour you might want to jump to chapter 4 which provides a quick start. Onthe other hand, the following sections will show you many newand exciting features of JCHESS.

2.2 Solids

As soon as natural systems are involved, precipitation and dissolution reactions become essential.Minerals are quite efficient buffers for important parameters such as pH and Eh. They also actas a source or sink for numerous elements and provide surfaces on which other ions may sorb...Solutions oversaturated with respect to a colloidal phase will precipitate out the colloids, creatingtremendous amounts of surfaces available for surface complexation reactions. Precise modellingof the behaviour of solid phases is therefore an essential concern in CHESS.

The second panel of the JCHESS, illustrated by Figure 2.3 provides several options which helpthe user to fine-tune precipitation and dissolution behaviour of minerals and colloids.

12

2.2. Solids

FIG . 2.3– Solids. Panel with options to fine-tune precipitation and dissolution beha-viour. The lower part shows species for which reaction kinetics have beenset.

2.2.1 Precipitation/dissolution at equilibrium

By default, CHESS assumes thermodynamic equilibrium with respect to solid phases and preci-pitation is enabled for all minerals. Hence solids precipitate only and exactly at the point wherethe Ion Activity Product (IAP) equals or exceeds the solubility constant (which is the inverseof the solid’s formation constant). Precipitation and dissolution of minerals and colloidal matterchanges the solution characteristics, notably the pH. Besides immobile minerals, CHESS also al-lows the precipitation and/or dissolution of colloidal (hence mobile) phases : an important aspect,since precipitation does not necessarily lead to hydrodynamic retention if the neoformed phaseis colloidal. The precipitation or dissolution process possibly changes the reactivity of the solidsurface, i.e. it creates or destroys functional groups or sorption sites in general. This processes isalso accounted for by CHESS.

By default, precipitation occurs as soon as the solution becomes over-saturated with respect toone or several minerals. However, CHESS allows to disable the precipitation process for allspecies (precipitation = disabled of all), for specific species (e.g.,precipitation =disabled of Hematite) or for a group (e.g.,precipitation = disabled of colloids).Click onof to select the species for which the precipitation reaction should be disabled.

Similarly, dissolution is enabled for all minerals, hence dissolution occurs as soon as the solutionbecomes under-saturated with respect to one or several minerals. In perfect analogy with theprecipitation options, dissolution can be enabled for all solids, one solid or a group of solids,thus allowing to adjust the behaviour of the system to specific needs.

For complex solutions, CHESS tries to find a stable mineral/colloidal assemblage by some kindof trial and error method. If you are curious to find out about this process, you might want to

13


take a look at the final report. Here you can find the code’s search path used to achieve finalequilibrium.

2.2.2 Kinetic control

It is well recognized that most precipitation and dissolution reactions arekineticallycontrolled.In that case, they are not necessarily in thermodynamic equilibrium with the solution. CHESSaccounts for reactionkinetics, generally mixed with the thermodynamic equilibrium approach.The following kinetic reaction for a solid (mineral, colloidal) species,S is used by the code :

dSdt

=

ApkpWp(Ωa−1)b if Ω ≥ 1

−AdkdWd(1−Ωc) f if Ω < 1(2.1)

where subscriptsp andd refer to precipitation and dissolution respectively,k denotes a kineticprecipitation constant in mol/m2/s, A is the volumetric surface area expressed in m2/m3, a, b, cand f are arbitrary power-constants, used to fit the law to experimental data,Ω is the 10-basedpower of the saturation index (IAP over Ks or IAP times the formation constant,K). As thelaw illustrates, the precipitation law is not necessarily the same as the dissolution law, even if itconcerns one single solid phase. The kinetic rate law is explained in detail in chapter 8 of thismanual.

W is a factor including reaction-sensitive species such as H+ or CO2(g) and may be differentfor precipitation and dissolution. Typically,W consists of one or more activities raised to somepower, the latter being a fitting parameter, for example :

W = [H+]m[O2(aq)]n. (2.2)

Note the inversion of the termsΩa−1 and 1−Ωc. If the term would not be> 0, use of the powercoefficients (b andc) may lead to an unexpected behaviour.

Kinetic control of a solid is obtained by setting the kinetics for the solid. Clicking onAdd...(right-clicking in the lower part of the panel opens a contextual menu) opens theKineticseditor which can be used to set the parameters used by the generalized kinetic law. Figure 2.4shows the editor which has, as all other editors used by JCHESS, three major fields : ’contexts’(currently only solid phases are proposed), ’items’ (the species) and the actual parameter confi-guration in ’selection’.

Select first a species from the list to which kinetic control will be added. You can either select aspecies from the item list or type the name of the species. It must exist though : if you want toadd a new species with kinetic control to the system, then addthis species in thedatabase panelfirst. The kinetic parameters are :

• Species :the name of the species

• Kinetic law :in theory, the type of kinetic law can be selected. For the time being, only one (default)kinetic law is available (eq. (2.1)).

14

2.2. Solids

FIG . 2.4– Kinetics editor. Panel which allows to set the parameters of the gene-ralized kinetic precipitation- and dissolution law.

• Reaction :by default, the reaction yields for precipitationanddissolution of the species. This is howkinetics are generally used. CHESS, however, allows to set alaw for precipitation onlyor dissolution only. If precipitation and dissolution both occur, but you want to set theparameters for each reactions to different values, then define two reaction kinetics for thesame species.

• Rate constant :the intrinsic kinetic rate constant in mol/m2/s (or similar units). The rate constant is theonly parameter which cannot be omitted. This parameter represents the precipitation aswell as the dissolution rate.

• Temperature :the temperature at which the rate was determined (calculated or measured). In absence ofa value, the temperature of the main solution is assumed. Note : this value does not changethe temperature of the solution.

• Activation energy :the Arrhenius activation energy, used to extend the rate constant to other temperatureranges. In absence of a specified value, the rate constant remains invariant whatever thetemperature.

• Nucleus :the minimum surface of the solid. Precipitation cannot start without an initial surface, sincethe overall reaction rate depends on the surface area. You may regard the nucleus surface

15


as very small, colloidal particles of unknown composition on which the mineral starts toprecipitate. If the specific surface of the mineral is known (this can be set as a speciesproperty in the main solution), then the specific surface is used as soon as the volumetricsurface (surface per liter of solution) exceeds the nucleussurface. This allows to include theeffect of changing surface area with concentration on the global kinetic rate. The nucleusis ignored if the reaction concerns only dissolution of the species.

• Saturation index :By default, precipitation occurs as soon as the saturation index exceeds the threshold of

0 and becomes positive. Using this parameter, precipitation occurs only as the saturationindex exceeds a higher threshold. By symmetry, dissolutiononly occurs as the saturationindex falls below the threshold. In absence of any specified value, an index of 0 is used.

• Dependence :Some species are well known to inhibit or catalyze the actualrate, such as H+. This tableallows you to set theW term, i.e. one or several species and corresponding ’powers’ onthe activities. The power ofΩ (coefficientsa andc of eq. (2.1) are introduced as follows :typeomega as species and set its power accordingly. Coefficientsb andd of eq. (2.1) areset by a similar trick : typeomega-1 as species and set its power accordingly, as illustratedin Figure 2.4.

Several examples for kinetic systems are presented in chapter 8, showing different aspects ofreaction kinetics with help of practical case studies.

2.3 Reactions

The next panel of the JCHESS interface definesreaction paths. A reaction path is the process ofbringing the main solution from its initial state to some other state. CHESS distinguishes threetypes of reaction paths, i.e.,titration, flushingandmixing. Figure 2.5 shows the general layout ofthe reaction panel. The following steps are needed to define areaction path :

1. set the reaction type (e.g., click ontitrate if you want to perform a titration with somevariable)

2. add one or several species using theadd... button of the reaction path panel and specifyone or more optional parameters

3. select the items or species you want to follow during the reaction in theOutput selectionpanel

You may set the number ofsamplesfor each type of reaction path. Each sample corresponds to anintermediary equilibrium state for which the selected items (defined in theOutput selection)are written to the output file. For titration- and mix reactions, the number of samples does notmodify the results, it merely defines the resolution of the graph you make of the selected items.The result of a flushing reaction path, however, is a functionof the number of steps, as outlinedbelow. By default, the number of samples is set to 100.

CHESS will liberate imposed quantities as soon as a reaction path starts. For example, if youimpose an initial pH of 6 in the main solution and you titrate with NaOH, the pH will change

16


FIG . 2.6– Titration editor, used to select species or variables for titration-likereaction paths.

tion to many other variables as well such as pe, time, temperature, solvent mass (H2O)... itis even possible and quite straightforward to titrate the main solution with colloids, organicor minerals. You can add these variables to the titration table using theadd... button. Thisopens theTitration editor.The titration editor lets you select variables to titrate with. Figure 2.6 shows a screen shotof the editor. Very similar to the concentration editor, this specific-purpose editor has threemajor fields :

– ContextsAll variables are classified in terms of contexts, which helps in finding species or

items faster than building one long list of thousands of items. JCHESS distinguishes 6different contexts, which are identical to the contexts of theConcentration editor,described in the previous section, except for thevariables context which now in-cludestime andtemperature as well. Indeed, kinetic or thermal reactions are mo-deledas if it were titration reactions, i.e. you bring the variable from an initial stateto a specified end-value ;

– Itemsall the items available for the chosen context. The item lists are fundamentally iden-tical to those described by theConcentration editor, except for thevariablescontext which is significantly extended.

– SelectionThis field allows you to set the end-value of the variable. Thestart value is defined

by the main solution, or 0 if undefined.

CHESS allows for multiple titration reactions. For example, you may want to bring thepH from 3 to 12 at the same time as you bring the Eh from−100 to 100 mV. However,this will disconnect the two variables (and Eh is strongly related to the pH). As a rule,multiple reaction paths should be handled with a maximum of precaution. It is not possibleto combine a titration reaction path with e.g. a flushing- or mixing reaction path.By default, linear steps from the initial state to the final state are used. You may changethe default behaviour tologarithmic stepsvia the check box at the left of the panel (see

18

2.3. Reactions

Figure 2.5). Note that some reaction paths reactions do not allow for logarithmic steps (i.e.,temperature) and the option is ignored.

• flushFlushing is a process where, at each step, part or all of the mother solution is removed andreplaced by the solution defined as theflushor secondarysolution. The secondary solutionis defined in a similar manner (i.e. using aconcentration editor as the main solution(seeMain solution, section 2.1).

The volume ratio of the main- and flush solution is relevant : for each sample, a fractionor the total volume of the mother solution is replaced by the flush solution. The size of thefraction is defined by the ratio of the main- and flush solutions and the number of samples.Let v be the volume fraction replaced at each step, then

v =flush solution volume

main solution volume×N(2.3)

whereN is the number of samples.If the flush solution contains minerals, and you do not want toadd these minerals to themain solution as part of the flush solution, selectmobile-fraction. Selecting mobilefraction excludes minerals from being included in the fraction replacing the main solution.Figure 2.7 illustrates the reaction panel configured for a flush reaction path.

Note that the process of flushing will change the pH or other imposed activities, concentra-tions and fugacities. It maintains the total concentration: the principle of mass conservationis guaranteed during the reaction path. To fix a variable suchas pH, specify the variable toneverliberate its activity.

• mixMixing is a process where, at each step, the main solution ismixedwith a fraction of thesolution defined as themix or secondarysolution. At the end of the reaction (last sample),both solutions are entirely mixed. The solution will be mixed with the main solution, andis set in exactly the same way as the main solution is defined (seeMain solution, sec-tion 2.1).You may set the volumes of the main- and mix solutions, hence setting the mixing ratio.By default, both solutions are set to 1 liter hence, after mixing, the main solution is 2 liter.The optionmobile-fraction is useful if the mix solution contains minerals which shouldnot be mixed with the main solution.

Note that mixing two solutions will change the pH or other imposed activities, concentra-tions and fugacities. To fix a variable such as pH, Eh, fugacities and imposed concentra-tions, specify the variable toneverliberate its activity.

• disabledCheckdisabled if you want to disable the current reaction path. Once disabled, CHESSonly calculates the equilibrium state of the main solution.

In CHESS, reaction paths require a selection of items which will be used for graphic outputpurposes. For each sample, the selected items will be written in columns to a file (CHESS.res

19


FIG . 2.7– Example of a flush reaction path configuration. The flushing solutioncontains a Ca-Na-Cl water type at a pH of 7.

20

2.3. Reactions

by default) which is used by graphic software. The item usually stands for the concentration ofspecies, but other items are possible as well. A unit may be specified, otherwise the default (SI)unit of the selected item is assumed. You can add items to the selection using theadd... button.This opens theOutput editor, described in detail below.

The order of selected items defines the default settings of the graph (all items but the first oneare plotted as a function of the first item). The safest approach is therefore to define e.g. thesample number as the first item, especially for flush and mix reaction paths. This order, however,is rapidly and easily changed usingJPlot’s options. Also, you can change the order of the itemsin the output-editor : right-click on one of the rows to move it upwards, downwards, to the top orto bottom :

2.3.2 Output editor

This editor lets you select species or items which are monitored during the reaction path : foreach sample, the selected items will be evaluated and written to a file (CHESS.res, by default)which will be visualized byJPlot (see section 2.8). More precisely, a graph will be drawn usingthe first selected item as the X-variable, the other selecteditems as Y-variables. Hence if, forexample, the reaction path involves a simulation of mineralprecipitation as a function of pH,a useful selection would be the pH andminerals, i.e. all minerals which exist or have existedduring the reaction. Figure 2.8 shows a screen shot of the editor. CHESS will give a warning ifno item has been selected for output.

As the concentration editor, this specific-purpose editor has three major fields :

• ContextsJCHESS distinguishes many different contexts. Thevariables context contains not onlypH, Eh and pe, but also many other items which are useful to monitor, e.g., time, tem-perature, the sample number, (sample), the ionic strength, the total mineral volume, etc.Thegroups item allows to select groups of items, such asminerals, colloids etc. Thisoption is very useful if we do not know, before hand, which mineral will precipitate duringthe reaction path. Selecting a species from theaqueous-species, colloids, organicsorminerals contexts allows to select the concentration of specific species. Selecting a gasdisplays the fugacity of the gas.

It is possible to select the mineral fraction of some basis species. For example, the mineralfraction of Ca2+ displays the sum of all Ca2+-containing minerals. Similarly, CHESS pro-

21


FIG . 2.8– Output editor. Here we define a selection of species or variables for gra-phical output.

poses to display the colloidal, organic, aqueous or fixed (e.g., bound by some surface site)fraction of a specific element.

There are two possible solutions if you want to follow the saturation indices of solid speciesduring a reaction path. First, you might want to selectall the saturation indices (contextvariables, itemsaturation-indices). However, this list might become quite long andfilled with minerals which are irrelevant. The contextsaturation-index is much morespecific, since it displays the saturation index of the selected solids only.

For minerals, colloids and organics, CHESS allows to selecttheir surface potential andsurface charge, provided the solid contains reactive sites.

• ItemsAll the items available for the chosen context. You can select an item from the list, or typethe item manually.

• SelectionThis field is used to manually add an item and/or to set the unitof the selected item. Someitems cannot be selected from the list, you have to provide them manually. For example,you might want to display all species containing the substring ’Zn’ in their name. Thisis readily obtained via the selection*Zn* (careful though : Zincite willnot be included).Similarly, A* selects all species starting with capitalA, *A selects all species ending withA. No other ’wild characters’ are recognized yet.

2.4 Sorption

This panel of JCHESS provides an ensemble of useful handles to fine-tune sorption reactions.Here one can set the type of solid-solute reaction, different electrostatic models andKD values.The panel is displayed by Figure 2.9.

22

2.4. Sorption

FIG . 2.9– Sorption. Panel with options to fine-tune the sorption models and/or to setKD values.

2.4.1 Thermodynamic sorption reactions

Thermodynamic sorption in CHESS is governed by more or less classical reactions with surfacesites. Two types of the mass-action law are recognized by CHESS, the classical (default) lawand a slightly different approach, referred to as the Amended Surface Complexation Approach(amended). Let us illustrate the difference with the example of cadmium complexation by a silicasurface :

Silica OH+Cd2+ Silica OCd+ +H+

which corresponds to the mass-law equation :

[ Silica OCd+] = K[ Silica OH][Cd2+][H+]−1

where [] denoteactivity of the quantities. Religious wars are ongoing on the exact meaning ofsite-activity : in the present context, we prefer to silently ignore this cumbersome issue and sim-ply assume that the activity equals the concentration multiplied by the electrostatic correctionterm (or by 1 in case of a zero net-change in surface charge). Note that you may disable electro-static correction all together or choose between the constant capacitance, double or triple layermodels.

The above equations are based on a crude hypothesis : they aremono-dentate, 1 :1 stoichiome-tric reactions. Imagine that your experimental data cannotbe modelled according to the above

23


reactions. One of the first steps to improve the modelling results is to consider other types ofreactions, such as a bidentate complexation reaction :

2 Silica OH+Cd2+ ( Silica O)2Cd+2H+

Here, Cd2+ is complexed bytwo hydroxylic groups present on the same surface. The classicalapproach would suggest the following mass-action law :

[( Silica O)2Cd+] = K[ Silica OH]2[Cd2+][H+]−2

The power of 2 used by the uncomplexed site-activity has an important effect on the overallcomplexation affinity with respect to Cd2+ ! An even much more pronounced effect is obtainedfor charge-neutralization reactions with e.g. tri-valentcations such as Am3+ and Eu3+.

Theamendedapproach changes the reaction affinity by assuming a reaction with only onesite,albeit composed of several hydroxylic groups. The approachtherefore uses a modified mass-action law : instead of the previous law, the amended law usesa site-density which is only halfof the functional group concentration :

[( Silica O)2Cd] = K[ Silica OH][Cd2+][H+]−2/2

As one may expect, the overall result is an entirely different reaction affinity. Note that it isnot possible to use the amended model for a reaction for whichthe log(K) has been definedpreviously for the default model. No database is currently available for this model. Nevertheless,disposing of this possibility appears to be a great help in modelling experimental data which cannot be interpreted using the more common conventions.

2.4.2 Distribution coefficients

The distribution coefficient (Kd) is often used to quantify the solid-bound fraction of a solute.Strictly speaking, the approach does not belong to the thermodynamic framework, since it fear-lessly violates the fundamental principle of mass conservation. This is readily illustrated, the dis-tribution coefficient being defined by the ratio of bound overfree solute. Hence, taking Eu(III)as solute, we have

free-Eu(III) bound-Eu(III)

and the corresponding reaction

Kd =bound-Eu(III)free-Eu(III)

We generally assume (and this is one of the vague aspects of theKd approach) that theboundfrac-tion only concerns Eu(III) actually sorbed by the solid, hence excluding (surface-)precipitation.CHESS uses a dimensionlessKd constant (molal over molal), instead of the more commonlyencountered unit of m3/kg (solution volume per unit mass of soil). It is readily verified that theconversion of this parameter to the dimensionlessKd of CHESS involves the soil density (ρs)and the porosityω :

Kd[−] =ρs

ωKd[m3/kg]

24

2.5. Database

Often, aKd value is obtained from a batch experiment whereω has no meaning. In that case, itis the solid-liquid volume ratio which should be used instead of ω :

Kd[−] =Vsρs

VlKd[m

3/kg]

The major advantage of the CHESS convention ofKd is that the value is slightly more intrinsic,hence one may expect to obtain reasonable simulation results even when using a solid-liquidratio different than the one used by the experiment.

This said, how do we actually set theKd of a species ? JCHESS provides a small specific editorwhich lets you select a basis species (only basis species canbe selected since theKd uses totalconcentrations only) and the appropriateKd value. That’s all there is : CHESS creates a (virtual)site namedKdSite-M which represents the sorbed fraction of metalM. A complete example ofa simulation withKd values is provided in Chapter 6.

2.5 Database

TheDatabase panel forms a gateway to the database. It lets you select the database, exclude orinclude species or groups of species, change properties of species such as the thermodynamicformation constants and even define new species. The powerful analyzer allows to learn moreabout specific species of the database : it shows the reaction, composition, formation constants,comments and references.

Figure 2.10 shows the panel, which is divided in several parts. Each part is explained in detailbelow.

• Current databaseThe path and name of the current database is displayed here. It is possible to open (and

even to edit) the database via the button at the right of this field. A more powerful wayto view database entries is theAnalyzer, as outlined below. TheReload button allowsto reload the current database in memory : this is necessary if you meanwhile changedthe current database. TheChange button opens a file-chooser and allows to load anotherdatabase. It is quite well possible to change the database for your current system : care mustbe taken, however, to use a database which actually includesthe species of the solution(s).If they do not exist, JCHESS will complain and disable the species.The Analyze button opens the Database Analyzer. New since version 2.0, the analyzerprovides a user-friendly and clear mean to obtain information from the database. It allowsyou to select a species, to illustrate its properties and reaction which forms the speciesas well as the corresponding thermodynamic constants as a function of the temperature.Moreover, it shows comments which concern the species and, if existing, references toarticles, books or authors. Figure 2.11 illustrates the analyzer for colloidal hydrous ferricoxide (known as>HFO by the database).The analyzer obtains all the information present in the current database. Many databasesdo not (yet) include comments, references, temperature dependencies or other properties

25


FIG . 2.10– Database. Panel through which the user interacts with the database.

FIG . 2.11– Analyzer. Analyzing Hydrous Ferric Oxide.

26

2.5. Database

and these are thusly not displayed by the analyzer : don’t blame the messenger... it’s allin the database. With these new tools available, the CHESS development team currentlyworks at new database formats including as much as possible relevant information for eachspecies. As these updated databases become available, theywill be uploaded to the CHESSweb site (http ://chess.ensmp.fr/databases.html).

• Sloppy readingBy default, CHESS will not allow errors such as charge-imbalances. Check sloppy readingif you want CHESS to ignore such errors.

• Exclude speciesCHESS provides two means to reduce the current database without actually editing the

database. The first mean is to exclude groups of species, i.e.minerals, colloids, organics,surface sites, etc. Since this option is used very often, JCHESS allows to simply check thegroup which should be excluded from the system. Exclusion ofgroups of species (e.g., allminerals) combined withinclusionof specific minerals is a straightforward and easy wayto reduce the database.

The second mean is to specifically select the species using the Exclude list. Clicking onEdit... opens a simple editor which lets you select one or several species (or groups ofspecies) which will be excluded from the system. Use this

• Include speciesClicking onEdit... opens a simple editor which lets you select one or several species (orgroups of species) which will be included. This feature is useful only in combination withexclusion of groups of species. Figure 2.10 shows an examplewhere all colloids, organicsand sites are excluded, except for colloidal silica (>Silica). This way, even the mostcomplete databases can be used to model simple laboratory solutions. Note that speciesalready defined in the system (e.g. set with a certain concentration in the Main solution)will not be excluded.

• (Re)define speciesIt is possible to modify one or several properties of a species, such as thermodynamic

constants, without editing the database. Modifying thermodynamic constants is particu-larly useful when fitting sorption data with CHESS. Note thatCHESS uses polynomials todescribe the log(K) as a function of the temperature.

Clicking onAdd... (or right-clicking in the table area) opens an editor closely related tothe Analyzer, except that each field is editable. You can takean existing species and modifyone or several properties (even it’s name) or entirely rewrite a new species. Note that thisspecies only exists for the current system, the underlying database is not modified.

To modify thermodynamic constants, double-click on the constant which enters the tablecell in editing mode. To validate the modified value, press ’Enter’, then click onApplywhich closes the editor.

27


2.6 Output

The simulation output is viewed and report-preferences areset via the output panel. The si-mulation results are written to two files :CHESS.outwhich is a general report providing theequilibrium results and a number of characteristic and useful parameters. This file can be viewedusing theView... button which pops up a simple viewer (see Figure 2.12). This viewer has beenadded for convenience : any editor can be used to view the file since it contains plain ASCII text.

If a reaction path is demanded, the general report also resumes the final state of the solution. The’output selection’ (defined in theReaction panel) is written as columns to a second file, namedCHESS.res. Note that this file is used by JPlot. Both filenames can be changed in theOutputpanel (see Figure 2.12).

The depth of detail of the general report can be altered usingtheOptions :

• fullprovides a long report, displaying all species used for the calculations and concentrationsdown to 10−300, saturation indices down to−300 as well as other complementary infor-mation.

• defaultprovides a default report (sufficiently detailed for most purposes).

• Species-cutoffsets the value of the lowest concentration displayed in the report. By default, CHESS cutsthe list at 10−25 molar (10−300 for a full report).

• SI-cutoffsets the value of the lowest saturation index displayed in the report. By default, CHESScuts the list at−3 (−300 for a full report).

• Report samplesets the sample number which will be displayed in the report.This option is useful for

reaction paths only. By default, the report summarizes the initial equilibrium state of themain solution, the initial state of the secondary solution (in case of a flushing or mixingreaction path) and the final solution state. Intermediate samples can be added to the generalreport (CHESS.outby default) by setting the sample number. You can set one (e.g., 43),several (e.g., 43, 45, 47) or all samples (*).

Note : you must re-run CHESS in order to take into account any changes made to the reportoptions.

If activated, theResume file resumes the solution properties at the end of the simulation (e.g.after a reaction path) in valid CHESS-script format. Reloading this file allows to pick-up thefinal solution to perform new reactions with.

28

2.6. Output

FIG . 2.12– Output. Here you can view the simulation results numerically, modify theoutput layout and set the resume file.

29


2.7 Piper

Hard-core hydrochemists don’t eat quich and display water analyses with Piper diagrams. Buteven novice geochemists might appreciate to see their analyses being characterized in terms of aPiper diagram. JCHESS is the only geochemical speciation tool which offers this possibility, andwe will briefly outline the nitty-gritty of this option.

To activate the option, go to thePiper panel andenablethe creation of piper data. At this stage,the model will create, at each run, a specific data file (named CHESS.ppr, by default). The data iscreated from the solution defined in themain-solution panel. JCHESS automatically pops upthe diagram after a run, or update an already displayed diagram. Figure 2.13 illustrates the Piperpanel and the diagram in which we plotted the solution characteristics of a mineral water :

FIG . 2.13– Piper diagram. Screen shot of the piper panel and a diagram with themain solution plotted.

30

2.7. Piper

pH = 7.0total Cl[-] = 31.2 mg/ltotal Ca[2+] = 555 mg/ltotal Mg[2+] = 110 mg/ltotal Na[+] = 14 mg/ltotal SO4[2-] = 1479 mg/ltotal HCO3[-] = 403 mg/l

Note that, by default, the solution TDS is also displayed as acircle around the data point, withits radius proportional to the total TDS.

Similarly, we can display thesecondarysolution (i.e. the one used for mixing or flushing) in thepiper diagram. Consider two liters of a second solution which will be mixed with one liter of theprevious mineral water :

total HCO3[-] = 1254 mg/ltotal Ca[2+] = 21 mg/ltotal Na[+] = 721 mg/ltotal Mn[2+] = 34 mg/ltotal Mg[2+] = 77 mg/ltotal Cl[-] = 684 mg/ltotal SO4[2-] = 41 mg/ltotal SiO2(aq) = 66 mg/l

The solution characteristics of this solution can be added to diagram : simply check thesecondarysolution option. Even the entire reaction path can be displayed : check the option and re-runCHESS. Figure 2.14 displays the initial solution, the mixing solution as well as the intermediatesolutionsduring the mixing process (samples). The TDS has been omitted for this example.

The Piper panel provides a wide spectrum of graphical settings to change the plot’s appearance.Most of the settings are common with JPlot, the graphical front-end of JCHESS (see 2.8). Amongothers, you can load other data files, provided they are formatted according to the piper diagramdata format as defined by CHESS. The piper data format is simple and contains only one line persolution, i.e.,

77.0941 20.7177 3.14005 76.7698 0.00263522

The numbers are the total aqueous percentages of respectively Ca, Mg, Cl, SO4 and the TDS(kg/kg). To obtain the percentages of e.g. Ca, the total aqueous concentration of Ca is divided bythe sum of Ca, Mg and (Na+K). Idem for Mg : the remaining fraction therefore corresponds toNa+K.

Data files are easy to construct. Each solution can be given a text label used by JPlot to identifythe solution in the legend, e.g. :

# row 1: Clear-water spring

31


FIG . 2.14– Piper diagram showing the initial solution, a sequence (samples) of inter-mediate solutions and the mixing solution.

# row 2: Black Hill’s main source

77.0941 20.7177 3.14005 76.7698 0.002635222.47821 15.2211 49.1016 1.77037 0.00297294

Accordingly, the Piper diagram feature of JCHESS can be usedeven without CHESS, but as agraphical package for specialized geochemical purposes.

2.8 JPlot

JPlot is JCHESS’ main plotting package, wrapped by the JCHESS interface to graphically dis-play the results of a reaction path calculation. JPlot is also available as a standalone applicationand aims at providing quality graphs : it is installed (a bonus) besides JCHESS on your computer.JPlot gives the user full control over line styles, colors, labels and text, drawing styles, axes, tics,legends and handles more than one datasets (files). Furthermore, JPlot is intuitive, easy to use,fast and versatile : JPlot provides common 2D plots but also Piper and activity-activity diagrams.

Still, it is quite easy to export the file containing the reaction path results to other graphic softwareor to your favorite spreadsheet program. The structure of the datafile is classical, i.e., columnswith a short description in the header. For example, the datafile can be used by the powerfulgnuplotplotting program without any post-processing.

JPlot’s toolbar can be detached from the application, or placed at other sides of the JPlot panel.Pressing one of the toolbar buttons leads to specific actions:

32

2.8. JPlot

clears the current system, resetting the graphic parameters to their default values.

opens a datafile. JPlot looks, by default, for data files with the .resand/or.datextensions.The datafile must have a valid format, i.e., at least two columns with valid values. The nameof the columns are defined in the header of the file : look atCHESS.resfor an example.In absence of a header, JChess names the columns by column number, i.e.,column 1,column 2, etc. ;

removes the datafile from the current system ;

shows the graph corresponding the the current X, Y selections ;

provides full control over the axes settings of the graph, including colors, bounding boxes(and fill-colors), tics, the background grid and more ;

scaling of the data on the X- and Y axes, logarithmic scaling,manual or automatic scaling ;

label and, more generally, text within the graph. JPlot allows to set specific labels (X-label, Y-label and title) but also unspecific, random labels. They are defined here and thendragged (using the mouse) to its final location.

the legend is located in the upper-left corner of the graph, but can be dragged (using themouse) to any place inside or outside the graph. Legend settings such as font and interline-spacing can be modified.

2.8.1 Data panel

If CHESS is run and a reaction pathway is set with a proper output selection (see section 2.3),JPlot’s main panel contains two main areas :

• the left-hand side contains a list of all items selected for output. You can click on an itemand set it as either the X-column, either one of the Y-columns.

• the right-hand side contains the columns actually selectedfor the graph. By default, JPlotassumes that the first column are the X-values and all the other columns are Y-values, butyou may select it differently.

Figure 2.15 illustrates the data panel for the titration of asolution containing cadmium (we se-lected pH and *Cd* inµmol/l). Note that not all species have been selected for the graph, butonly the relevant species. The approach is to make a rough selection in theReactions panel,then to refine the selection (as a function of the results) in this panel. Colors, line and point stylesare set via a single click on thestyle cells of the selected Y items. This opens the style editor,illustrated by figure 2.16. The style chooser proposes many options, which are self-explanatory.It is possible to change the text used for the legend, or to hide the legend of this curve. Anotheroption is to multiply the data of this column with a value, or to add a value : the usefulness ofthis option shows up when you need to fit the model results withexperimental data.

Click on the graph icon ( ) to display the graph (Figure 2.17). Note that the graph is auto-matically updated after each run of CHESS. Manually updating is possible via one of the buttons

33


FIG . 2.15– JPlot. Main panel with selectable columns at the left, selected columns atthe right.

in the graph’s toolbar. By default, JPlot places the legend in the upper left corner of the graph.You can drag the legend to any other place of the graph. The graph’s toolbar contains buttons toclose the graph window, to reload the graph, to export the graph (currently only the JPEG formatis supported) and to print the graph. You can print the graph to either the printer (default) orto a postscript file. A printer chooser helps you to select a printer and to change, if needed, thesettings.

And here ends the tour through Speciation Mode. The next chapter will guide you through ano-ther model of JCHESS, the diagram mode. The remaining chapters go beyond the interface andintroduce the model from a geochemist, scientific view point.

34

2.8. JPlot

FIG . 2.16– JPlot. Plot, line and point styles chooser.

FIG . 2.17– JPlot. Click on the graph icon of the main panel to show the graph. Thiswindow will stay open and is updated after each run of CHESS.

35

Chapitre 3. A walk through Diagram Mode

Diagram mode provides a handle to activity diagrams, such asPourbaix, solubility and mosaicdiagrams. Originally developed by Pourbaix for the field of aqueous corrosion of metals, Pour-baix diagrams have proven to be very useful in hydro-metallurgy but also for geochemistry. Theyprovide predominance fields of a specific species as functions of pH and redox potential (Eh orpe).

Activity diagrams are an extension of Pourbaix diagram in the sense that they accept the acti-vity of whatever species on X- and Y axes. They thusly providea quick, albeit approximativeoverview of the dominating species with respect to the axes variables.

Solubility diagrams are, in turn a special case of activity-activity diagrams, since they definethe activity of the diagram species as the Y-axis variable. An increasing activity of the speciesleads, at some point, to oversaturation with respect to minerals formed by the diagram species,therefore defining the solubility-limited area.

With JCHESS, the backlog of work generally demanded to manually construct a diagram isreduced to a minimum. Moreover, the robustness of the mathematical approach allows to buildgraphs which are too awkward to build by hand, such as mosaic diagrams implying multipleinterdependent species. JPlot has been amended to produce high quality diagrams, ready for usein reports and scientific papers.

By default, JCHESS starts inSpeciation mode (see Chapter 2). SelectDiagram mode in themodels menu to switch. Launching JCHESS with the-a inverts the rule, and starts the model inDiagram mode.

3.1 Water reduction and oxidizing

Water can act both as an oxidizing agent (reduction to H2) and a reducing agent (an attendantproduction of O2). Since thermodynamic equilibrium requires that a chemical species should notreact with the solvent through a redox reaction, only the region where the solvent remains stableshould be considered. By default, JCHESS excludes the regions where the solvent, H2O, tendsto dissociate : these fields are dominated by either O2(g) and H2(g).

In practice, an extra energy supply (i.e. anovervoltageor overpotential) is often needed to ob-tain water dissociation, due to kinetic processes. This is always found when gases are involvedin the redox processes. To visualize activity fields beyond the theoretical stability field of wa-ter, JCHESS provides several options. Checking thetransparent limits option still showsthe theoretical limits but displays the activity diagram beyond the stability field. Checking theignore limits option will not plot the stability limits at all.

37


FIG . 3.1– Pourbaix diagram for Pb with an activity of 10−4.

3.2 Pourbaix diagrams

Speciation of lead

In order to illustrate the workings of diagram mode, let’s consider a simple Pourbaix diagram forlead (Pb). Construction of the diagram takes exactly three steps :

1. set the X-range, e.g. pH from 0 to 14

2. set the Y-range, e.g. Eh from -0.8 to 1 V

3. add the activity of Pb to the system

Hit the run button (building an activity diagram requires a CHESS run) to display the graph.The result should be something very similar to the screen shot displayed by Figure 3.1. Theactivity diagram is built with a fixed activity of Pb2+, used to speciate Pb as a function of theaxes variables, pH and Eh.

The diagram graph, thus created, shows a number of fields. Fields with labels H2(g) and O2(g)denote domains where H2O dissociation reaches such a level that common thermodynamic equi-librium laws in H2O are no longer valid. All other fields denote areas for which aspecific Pb

38

3.2. Pourbaix diagrams

FIG . 3.2– Speciation diagram of 10−4 molal of Pb and an Eh of 0.1 V.

species dominates. We distinguish aqueous species Pb2+ under acid conditions, PbOH+ andPb6(OH)84+ for a pH range of 7.5 to 9 and Pb(OH)3

− at high pH, but also several redox-sensitivesolids : Pb, Litharge (Pb(II)), Minium (a mixture of Pb(II) and Pb(IV)) and Plattnerite (Pb(IV)).

How does an activity diagram compare with a speciation graph, as produced in speciation mode ?Remember that, in activity mode, all equations are evaluated directly without solving mass-balances which is only possible when knowing the activity ofthe basis species. Figure 3.2 showsthe result obtained speciation mode for a total concentration of 10−4 molal of Pb and a constantEh of 0.1 V. The graph roughly corroborates the diagram findings : Pb2+ predominates at lowpH, Pb(OH)3− at high pH, but in between, the dominating species are slightly different. Themost important difference, however, is that in speciation mode we also quantify non-dominatingspecies. For example, an activity diagram predicts domination of Litharge in the pH range of 9– 11.5 : in reality, Litharge dominates for a smaller pH rangeand several other species co-existat quite significant concentrations, such as Pb3(OH)42+, Pb(OH)2(aq)and Pb(OH)3−. Althoughuseful for a first estimate, one should always keep in mind thesimplifying assumptions of anactivity diagram with respect to a full speciation calculation.

Pourbaix diagrams have, classically, Eh on their Y-axis to specify the redox state. Nothing wi-thholds you, however, to use another redox-defining parameter such as pe or oxygen (O2(g) orO2(aq)). Figure 3.3 shows the speciation of arsenic as a function ofpH and O2(aq). Note that thetransparent water stability limits options was used. To produce reducing conditions,geochemical models typically require aqueous oxygen activities in the order of 10−30–10−80.

Adding an auxiliary activity : mosaic diagrams

Let’s extend the example of the previous section by adding another reactant to the system,HCO3

−. Lead tends to form quite strong carbonate complexes such as(hydro)cerussite. To eva-luate the effect of carbonate, we have to set the activity of HCO3

−. We have to choice, howe-

39


FIG . 3.3– Activity diagram of 10−6 molal of arsenic as a function of pH and oxygenactivity.

ver, between specifying atrue activity(specifieractivity) or a total activity(specifiertotal).Using the one or the other often leads to quite different results : illustrating and explaining thedifference is the main objective of this section.

We will set the activity of HCO3− to 10−3, with theactivity keyword. This will fix the activityof species HCO3− to this value, whatever the pH and Eh. In other words, we assume that theactivity is independent of any of the axes variables —a quitecommon assumption in activitydiagrams. The result is shown in the upper diagram of Figure 3.5.

Obviously, the result is not accurate since carbonate speciation strongly depends on pH. Up to apH of 6.3, species CO2(aq) dominates, while for pH> 10.3, species CO32− is the predominantcarbonate species. Use of a constant activity for a wide pH domain is one of the drawbacks (anddangerous pitfalls) of diagram mode and Pourbaix diagrams in general.

An approximate solution to this problem is to subdivide the diagram into several domains, thuscreating several sub-diagrams for each predominant carbonate species. This type of diagramsis sometimes referred to as amosaicdiagram, and is readily obtained with JCHESS using thespecifiertotal. JCHESS will speciate secondary species (such as carbonatein our example) whenspecifying a total activity of the species (see Figure 3.4).The second diagram of Figure 3.5illustrates the results for this case. To obtain the graph, JCHESS constructs a mosaic diagram. Itis possible to specify more than one dependent variables —the number of species which can beadded to the solution is unlimited.

As outlined in the previous, CHESS builds mosaic diagrams asa classic puzzle. Each piece cor-responds to a domain where a specific species dominates (CO2(aq), HCO3

−,. . .), and a speciationdiagram is solved for each piece. Then, all fields with commondominating species are merged,

40

3.3. Adding more generality

FIG . 3.4– Editor to select HCO3− as an auxiliary species, specified by itstotal activity

(see text) as opposed to its true activity.

a process nameddiagram purificationin CHESS parlance. For the curious : purification can beturned off via the preferences menu.

3.3 Adding more generality

Generic activity diagrams are readily constructed by JCHESS, since virtually all kind of speciescan be used on the axes. For example, a typical diagram can be obtained by using pH on theX-axis andcarbonateor nitrate activity on the Y-axis. Again, the user has the choice betweenimposing a true activity, or atotal activity on the Y-axis. Imposing a total activity allows tospeciate the Y-axis variable along the X-axis.

Solubility diagrams are a special case of the more general activity-activity diagram, as they usethe activity of the main species as Y-axis variable. Increasing the activity leads, at some point, tooversaturation with respect to mineral species formed by the diagram species, therefore illustra-ting the solubility of the species.

Let’s consider the case of Al at 50C. Aluminium is very unsoluble in natural systems andgenerally occurs in the form of Diaspore (AlOOH) or other, more complex minerals, in presenceof e.g. silicium, sodium and potassium. A solubility diagram of Al is constructed exactly thesame way we have seen for Pourbaix diagrams. We take, however, the activity of species Al3+

as the Y-axis variable while, at the same time, Al3+ is chosen as the diagram variable. Figure 3.6illustrates the solubility diagram in absence of any other species in solution. To clarify even morethe graph, the mineral phase has been highlighted as a gray area.

3.4 Improvement of graphical aspects of the diagram

Coloring surfaces

Each field of the diagram is provided as anitem in the JPlot panel. Clicking on the drawing-properties (left cell of the selected items) opens theLinestyle Editor, which can be used to

41


FIG . 3.5– Speciation of lead (Pb) with carbonate in the background. Top figure : si-mulation with an imposed activity of HCO3

− (specifieractivity). Bot-tom figure : simulation with an imposedtotal activityof HCO3

− (specifiertotal), leading to a mosaic diagram.

42

3.4. Improvement of graphical aspects of the diagram

FIG . 3.6– Solubility diagram for aluminium at 50C.

set not only the color of the lines used to draw the field, but also to set the background color.By default, JCHESS uses a transparent background. Press theleft-most button on the top of theeditor to activate the colored background and set a specific color.

Labels

JCHESS does its upper-best to appropriately place the labels with the species names. Puristmight prefer seeing the labels a bit more out-centered, rotated or simply removed. All flavors arepossible since JPlot allows to drag and drop the labels by mouse and to alter most label propertiesvia the Labels menu of JPlot.

Merging multiple diagrams

With JPlot it is quite straight-forward to merge two or more diagrams. The data of the diagramis stored in the general CHESS result file, defined in theOutput panel of JCHESS. This fileis namedCHESS.resby default, but can be set to any other filename. To save a diagram underanother name, changeCHESS.resto another name (e.g.diagram1.res). Run CHESS to actuallycreate the datafile.

Now build the second diagram and save this diagram under another name, e.g.diagram2.res.Select theJPlot panel, clear the system and loaddiagram1.resanddiagram2.res. At this stage,JPlot displays both graphs together. To achieve a more clearresult, you could consider modifyingcolors of, e.g., the labels and lines to distinguish both diagrams.

The remaining chapters go beyond the interface and introduce the model from a geochemist,

43


scientific view point.

44

Chapitre 4. Getting started

4.1 Running your first example

Most programmers have started their career with this simpleprogramme : a few lines of codewhich writes ’hello world’ to the screen. The modelling geochemist could start with a simi-lar one-liner, for example dissolution of quartz in pure water. Quartz dissolves partly, releasingSiO2(aq) —which acts as an acid. How much exactly of SiO2(aq) will be in equilibrium withquartz ? And what is the resulting pH ? CHESS might be able to tell you this with great preci-sion, as outlined below.

The most simple example CHESS can provide to you is to calculate the equilibrium state of purewater. Launch JChess, if not launched already. Since H2O is always present in the system (bydefault at a total concentration of 1 kg), nothing needs to beentered. The example only requiresto run CHESS (use the tool bar button).

The results can be examined via theOutput panel. By default, CHESS writes the results to anASCII file namedCHESS.out, which can be viewed with any text editor. JCHESS integratesasimple viewer, which pops up after pushing theShow... button in the output panel. Since readingthe final report is something we ask for so often, atoolbar button is provided on the toolbar toshow the output file immediately. The report contains much information, but for the moment wehave a quick look at only a small part of the file :

4 species were considered for the equilibrium calculations1 solids were considered for the equilibrium calculations

==============================================================================

calculating initial equilibrium......the system converged in 4 iterations...success!

Final equilibrium of the main solution:=======================================

pH : 6.99755ionic strength : 1.00603e-07temperature : 25 Celsiuselectrical imbalance : 9.775e-26 eq/lcarbonate alkalinity : 0 eq/lsolvent activity : 1solvent mass : 1000 gtotal dissolved solids : 1.8124e-09 kg/kgsolution mass : 1000 gsolution density : 977.5 g/lsolution volume : 1.023 liter

45


Aqueous species:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---OH[-] 1.006e-07 9.8339e-08 1.6725e-06 1.711e-06H[+] 1.006e-07 9.8339e-08 9.9116e-08 1.014e-07

The first block of the file resumes some important modelling assumptions (e.g. which versionof CHESS was used, what kind of activity model was use, which database, etc.). It also showsthat 4 non-solid species and mineral was considered for the calculations, namely H2O, H+, OH−,H2O(g) and ...ice (present as a mineral species in the database) which is, of course, undersaturatedat room temperature.

CHESS needed 4 Newton-Raphson iterations to converge to a final solution, which is reportedin the second block. Events during the calculation procedure are reported here as well, such asprecipitation and/or dissolution events.

The third block contains the simulation results corresponding to the main solution. Without muchof a surprise, the report presents an equal amount of H+ and OH− ions and therefore an equili-brium pH of 7. Internally, CHESS uses molalities (i.e. molesper kg of solvent) but the results arelisted in other commonly used units as well, such as molarity(moles per liter of solution), gramsper liter of solution or grams. The solution density is calculated as a function of the solutiontemperature.

4.2 Quartz in pure water

Consider a solution of pure water and 10 mg/l of quartz. Pressing theadd... button of theMainsolution panel opens the concentration editor (see Figure 4.1). The left-most list displays thecontexts : they are used to classify the different items, species and variables proposed by themodel and by the database. Quartz is a mineral, hence we select the contextminerals, whichfills the item list with mineral species. Select item ’Quartz’ and set the value at 10 mg/l and pressApply to add this item to the main solution composition. Push theDismiss button to close theconcentration editor.

And that’s all for the moment. Launch CHESS using the toolbarbutton, which will calculatethe equilibrium state of a solution which only contains 1 kg of water and 10 mg of quartz. Asoutlined before, the results can be examined in the final report available in theoutput panel orvia the specific button on the toolbar. This time, 12 aqueous species and gases are considered, aswell as 9 solids1. Let’s have a look at the main block of the report :

Final equilibrium of the main solution:=======================================

pH : 6.83594ionic strength : 1.45967e-07

1To see a list of all species considered by the run, select afull report in theoutput panel.

46

4.2. Quartz in pure water

FIG . 4.1– Building a simple solution using the concentration editorof JCHESS.

47


temperature : 25 Celsiuselectrical imbalance : 9.775e-26 eq/lcarbonate alkalinity : 0 eq/lsolvent activity : 1solvent mass : 1000 gtotal dissolved solids : 6.0254e-06 kg/kgsolution mass : 1000 gsolution density : 977.51 g/lsolution volume : 1.023 liter

Aqueous species:---------------- ---molal--- ---mol/l--- ---g/l---- ---grams---SiO2(aq) 0.00010016 9.7908e-05 0.0058827 0.0060181H[+] 1.4597e-07 1.4268e-07 1.4381e-07 1.4712e-07HSiO3[-] 7.6619e-08 7.4895e-08 5.7738e-06 5.9067e-06OH[-] 6.9348e-08 6.7788e-08 1.1529e-06 1.1794e-06H2SiO4[2-] 5.1684e-14 5.0521e-14 4.754e-12 4.8634e-12H6(H2SiO4)4[2-] 1.0851e-16 1.0607e-16 4.0564e-14 4.1498e-14H4(H2SiO4)4[4-] 2.5684e-25 2.5106e-25 9.5511e-23 9.771e-23

Minerals and colloids:---------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---Quartz 7.0026e-05 6.845e-05 0.0041128 0.0042075

Hence 10 mg/l of quartz leads to a pH of∼6.8 and about 60 % of the quartz has dissolved. Notethat with only one mineral in pure water, we obtain a system of12 aqueous species and 9 potentialminerals and colloids (with CHESS, species starting with> are considered colloidal). Since weare at thermodynamic equilibrium, all the solid species areunder-saturated as illustrated by thelist of saturation indices :

Saturation indices (down to -3) of solids:------------------------------------------colloid >Quartz 0mineral Ice -0.1387mineral Tridymite -0.1715mineral Chalcedony -0.2712mineral Cristobalite(alpha) -0.5505mineral Coesite -0.81mineral Cristobalite(beta) -0.994mineral SiO2(am) -1.286

CHESS handles much more complex systems, but the procedure is generally the same : composethe main solution, run CHESS and check the output file for the results. The next chapters providean in-depth and detailed introduction to CHESS, showing theeffect of different options andseveral reaction paths with help of practical examples.

48

Chapitre 5. Carrying on

Having learned to walk, let’s try a jog ! Several simple but practical examples will show you howto simulate the equilibrium state of laboratory systems or natural media. This chapter deals withspeciation, precipitation and dissolution of colloidal and mineral solids and the redox reactions.

5.1 Solubility of ferric iron

We will consider a laboratory experiment designed to study the solubility of ferric iron. Theexperimental solution is composed of 0.1 mol/l NaCl and 1 mmol/l FeCl3. The pH is adjustedto a value of 8 by adding sufficient amounts of NaOH. How do we translate this chemistry inmodelling parlance ?

We know the total amount of elements Cl and Fe, and we have a pH.But nothing is known yetabout the actual concentration of, e.g., species Fe(OH)2+ and there is uncertainty about the exactamount of element Na. One way to proceed is to introduce thetotal concentrationsof elementsNa, Cl and Fe calculated from the input concentrations of NaCl and FeCl3 salts. Furthermore,we use species Na+ to balance the solution electrically. The main solution then looks like this :

For simplicity, we exclude any colloids precipitation fromthe system : go to theDatabase paneland check ’exclude colloids’ which add colloids to the list of excluded species . Note that speciesFe3+ is not a basis (default) species for Fe : you will find Fe3+ in the list ofaqueous speciesinstead. Now run CHESS and open the report to view the results. Here is a snippet of the file :

49


calculating initial equilibrium......the system converged in 97 iterations

precipitating Hematite, saturation index = 17.8609...the system converged in 5 iterations...success!

About 97 Newton-Raphson iterations are needed to find a mathematical solution of the equi-librium state. Automatically the model decides to precipitate this mineral and recalculate theequilibrium state, this time in presence of the just-born solid phase. Then, much closer to the trueequilibrium solution, convergence is reached in only 5 iterations.

Another part of the reports shows the concentration and activity of the aqueous species, theminerals (and colloids) if present and a table with cumulative, total concentrations :

Aqueous species:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---Cl[-] 0.10427 0.10192 3.6135 3.6965Na[+] 0.10427 0.10192 2.3432 2.397NaCl(aq) 0.0011042 0.0010793 0.06308 0.06453OH[-] 1.2973e-06 1.2681e-06 2.1567e-05 2.2063e-05NaOH(aq) 1.3038e-08 1.2745e-08 5.0976e-07 5.2148e-07H[+] 1.2828e-08 1.2539e-08 1.2638e-08 1.2929e-08HCl(aq) 1.7379e-10 1.6988e-10 6.1939e-09 6.3363e-09Fe(OH)3(aq) 1.1332e-12 1.1077e-12 1.1838e-10 1.211e-10Fe(OH)4[-] 3.6512e-14 3.569e-14 4.4212e-12 4.5228e-12Fe(OH)2[+] 3.1077e-14 3.0377e-14 2.7298e-12 2.7925e-12FeOH[2+] 1.9808e-18 1.9362e-18 1.4106e-16 1.443e-16Fe[3+] 1.0654e-23 1.0414e-23 5.8159e-22 5.9496e-22FeCl2[+] 1.2956e-24 1.2665e-24 1.6053e-22 1.6422e-22

Minerals and colloids:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---Hematite 0.00051152 0.00050001 0.079848 0.081684

Total concentrations:--------------------- ---molal--- ---mol/l--- ----g/l---- ---grams---H2O 55.51 54.26 977.5 1000H[+] 1.3e-08 1.271e-08 1.281e-08 1.31e-08Na[+] 0.1054 0.103 2.368 2.422Cl[-] 0.1054 0.103 3.652 3.736Fe[3+] 0.001023 0.001 0.05585 0.05713

Since we asked to balance the solution using Na+, the solution is electrically neutral, and thefinal total Na concentration, given in the second table is about 103 mmol/l. The speciation ofchlorine and sodium is quite simple, whereas the iron chemistry is more complex. The protonactivity is, of course, precisely 10−8 (activities are shown only in the ’full’ report option, whichcan be set in theOutput panel. Note that, by default, the activities of the aqueous species are

50

5.1. Solubility of ferric iron

computed with thetruncated Daviesformula : it is possible to specify other models, such as theDaviesformula, and two formulae derived from theDebye-Huckeltheory. These can be set in theMain solution panel. Their theoretical background is discussed in detailin van der Lee (1998).

The choice of Fe3+ as our ’basis species’ was guided by a direct translation of the stoichiometryof the FeCl3 salt rather than by geochemical insight. From a numerical point of view, it was notthe most appropriate choice because it leads to a bad initialguess of the solution (we recall thatthe solution is obtained iteratively, hence the need for an initial guess). At equilibrium, iron isfound in the form of mainly hematite.

It is not difficult to help the model numerically by choosing amore appropriate set of represen-tative species. The use of amini-tableauis very helpful, as outlined in van der Lee (1998). Here,we reproduce a mini-tableau for our example, replacing Fe3+ by the thermodynamically morestable hematite (see Table 5.1).

The concentrations of our new basis are calculated by multiplying the stoichiometric coefficientswith the total concentrations corresponding to the recipe of that row and summing up all rowsof the column. Note that the total concentration of H+ is undefined : H+ is constrained by itsactivity (hence its total concentrationfloats). The species H2O forms automatically part of thebasis and does not need to be included in the basis. Accordingly, we can write the solution alsoas follows :

Running CHESS with this script leads to virtually identicalresults as the former script, but thenumerical behavior is quite different :


51


Na+ Cl− Hematite H+ H2ONa+ 1 0 0 0 0 0.1 mol/lCl− 0 1 0 0 0 0.103 mol/lFe3+ 0 0 0.5 3 −1.5 0.001 mol/lH+ 0 0 0 1 0 ?

H2O 0 0 0 0 1 55.6 mol/lTotal concentration 0.1 0.103 0.0005 ? 55.6 mol/l

TAB . 5.1– Mini-tableauused to evaluate the total concentration of a set of alternativebasis species.

...success!

Hence simply by changing the set of basis species, the model takes only about 8 iterations toconverge to the same results. The choice of the set of representative species is of no consequencesin our simple example, but it may become more important for systems including more complexmineral phases (especially alumino-silicates) and/or redox species. This problem, related to theway the mathematical solver of geochemical models works, stresses the usefulness of some basicknowledge of not only chemistry, but also of basic concepts in geochemical modelling.

5.2 Precipitation and dissolution

The previous example already illustrated how the model accounts for precipitation. Dissolutionand precipitation of solid phases (i.e. colloids and minerals) are important geochemical pro-cesses, omnipresent in nature. CHESS behaves fairly autonomous with respect to precipitationand dissolution, in the sense that solid phases are automatically precipitated or dissolved if nee-ded. In other words, the model assumes thermodynamic equilibrium for all species, includingsolids such as colloids and minerals.

However, precipitation and dissolution are time-dependent, kinetics reactions. Non-thermodynamicfactors such as nucleation (for precipitation) or surface availability (for precipitationanddisso-lution) are important to the reaction of solids (see e.g. vander Lee (1997) for a section on thismatter). In order to anticipate these phenomena, precipitation and dissolution can be disabled or”fine-tuned” for specific case studies, where thermodynamicequilibrium with some minerals isunlikely to be reached. The most simple tuning options are todisable or enable dissolution and/orprecipitation. More refined adjustment can be obtained by e.g. disabling precipitation or dissolu-tion of specific species only, modifying thermodynamic equilibrium constants, excluding solidsfrom the system, or accounting for kinetic rate laws. All these and related options are found intheSolids panel.

In order to demonstrate the possibilities with respect to precipitation and dissolution, let us consi-der another experimental system in which we introduce colloidal hydrous ferric oxide (HFO in

52

5.2. Precipitation and dissolution

short) at a concentration of 10 mg/l in a 10 mmolar backgroundelectrolyte (NaClO4) to initiatea batch sorption experiment. The main solution now containsthe following items :

pH = 5total Na+ = 10 mmol/ltotal Cl04

− = 10 mmol/lcolloid >HFO = 10 mg/l

We balance on ClO4− in order to assure electroneutrality. Although HFO may be perfectly stableunder laboratory conditions, it is thermodynamically lessstable than hematite. And indeed, run-ning CHESS shows us a total dissolution of the colloids to form a more stable mineral phase :


precipitating Hematite, saturation index = 4.1668...the system converged in 6 iterations...success!

As illustrated by the script, hematite immediately replaced >HFO, which is not exactly what wewant. One of the options to alter the thermodynamics is to exclude hematite from the system :simply add this species to the list of excluded species in theDatabase panel. However, themineralgoethite will precipitate instead, hence you still loose the hydrous ferric oxide colloids.We might consider to exclude goethite as well... but what if we want to add HFO colloids in asystem which contains also goethite ? The answer is : modify the precipitation and dissolutionbehavior.

The first option is to disable dissolution of>HFO, hence disabling precipitation of secondaryphases composed of HFO-forming species (principally Fe(III)). This is possible via the optionsprovided by theSolids panel : disable dissolution, then click onof which pops up a selectorand select>HFO :

At this stage, dissolution is disabled of only one species, colloidal hydrous ferric oxide. RunningCHESS on this system indeed keeps the colloids in solution. However, the solution is now de-prived of any trace of Fe(III), which is unrealistic. We assume that a fraction of hydrous ferriciron colloids will dissolve, albeit a very small fraction. Sometimes, e.g., for colloidal silica, thedissolvedcomponents can play an important role in the speciation of trace metals by acting ascomplexing agents.

Another option is to let the colloids dissolve (partially),but to prevent precipitation of newphases. Indeed, precipitation is kinetically controlled and often requires a super-saturated so-lution to trigger the precipitation process. As with dissolution, we have the choice to disable the

53


entire precipitation process or to disable precipitation of one or several solids. Disabling precipi-tation of specific species is different from excluding thesespecies : the former option only skipsthe test on the saturation state of the mineral, while keyword exclude eliminates a species fromthe database used by the run. Accordingly, disabling precipitation allows to examine saturationstates by means of the saturation indices.

Note : Although colloidal and mineral hydrous ferric oxides have identical formation constants,colloidal species will precipitate first. This choice of CHESS has been made by purpose, sincewe expect that colloidal phases are formed prior to minerals. One may dislike this choice andalter CHESS’ behavior by slightly changing the formation constants of one of the solids in theDatabase panel.

5.3 Redox disequilibrium

By default, CHESS assumes no redox equilibrium, since CHESShas its origins in modellingof laboratory systems where redox equilibrium is not likelyto be reached. Redox is enabled,however, as soon as the Eh or pe is set. But more sophisticatedmanagement of the redox stateis possible. The assumption of thermodynamic equilibrium in the case ofredox-sensitive speciesof a geochemical system is not always justified in natural waters, especially at low temperature.In addition, an Eh measurement may reflect the predominant redox couple, or at least the fastestreacting couple during the measurement, but it does not forcedly represent the redox potentialfor all the other redox couples. For each solution, a detailed management of the redox behavioris proposed. You can set the redox potential manually if, forexample, the Eh is unknown but youknow the concentration of one or several redox-sensitive species.

5.3.1 Pyrite dissolution

For example, we might be interested in knowing the equilibrium state of pyrite in pure water. Weintroduce 1 gram of mineral Pyrite in the main solution and enable redox reactions. Run CHESSand study the general report. Here is a part of it :

54

5.3. Redox disequilibrium

pH : 7.03168ionic strength : 1.82518e-07temperature : 25 Celsiusredox potential (Eh) : -0.22307 voltselectron activity (pe) : -3.7698electrical imbalance : 2.9325e-25 eq/lcarbonate alkalinity : 0 eq/lsolvent activity : 1solvent mass : 1000 gtotal dissolved solids : 5.7726e-09 kg/kgsolution mass : 1000 gsolution density : 977.5 g/lsolution volume : 1.023 liter

Aqueous species:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---OH[-] 1.0884e-07 1.0639e-07 1.8095e-06 1.8511e-06H[+] 9.3012e-08 9.0919e-08 9.1637e-08 9.3746e-08Fe[2+] 2.7486e-08 2.6868e-08 1.5005e-06 1.535e-06HS[-] 2.5328e-08 2.4758e-08 8.1886e-07 8.3771e-07H2S(aq) 2.2878e-08 2.2363e-08 7.6217e-07 7.7971e-07SO4[2-] 6.9536e-09 6.7972e-09 6.5296e-07 6.6799e-07

CHESS has calculated a redox potential (Eh) of−223 mV and a neutral pH. The redox beha-vior can be further fine-tuned. For example, analogous to precipitation and dissolution reactions,redox can be disabled for one or several species. The commandredox = disabled of can beused to prevent some specific species representing different oxidation states of an element to belinked to equilibrium reactions. Once a redox coupled has been disabled, CHESS considers thedisabled species as equivalent to an independent species.

This option is particularly useful when measurements show that, for some species, redox equi-librium is not reached. The disabled species has thus its ownmass balance and an independentconcentration, which must be given by the user. The coupled redox species are listed in appen-dix A.

5.3.2 Decoupled redox state

We carry on with another practical example. Groundwaters circulating in a crystalline fractu-red rock have been sampled in boreholes with a packer system,such that anoxic conditions arepreserved during the sampling phase. The fractures are filled with carbonates (calcite, siderite),pyrite, quartz and clay minerals. The groundwater samples are analyzed, revealing, mildly re-ducing conditions with an Eh close to zero. The redox couple Fe(III)/Fe(II) appears to be inequilibrium, conform to the Eh. Denitrification by pyrite oxidation, which is thermodynamicallypossible, does not occur for unknown reasons, although kinetics may explain part of this obser-vation. We use the CHESS model to verify the analysis. The rock water composition is resumedas follows :

55


Note the temperature, set to 12C. The observed redox disequilibrium state for nitrates areexplicitly taken into account by CHESS since both N2(aq) and NO3

− were specified, togetherwith the redox optionredox is disabled of NO3[-].

Running CHESS and examining the general output report showsthat the Fe(III)/Fe(II) coupleseems to be in redox equilibrium, controlled by the paragenesis siderite/Fe(OH)3. This is confir-med by the calculated saturation indices of these minerals,obtained when running CHESS withprecipitation disabled :

Quartz : 0.28Calcite : 0.12

Fe(OH)3 : 0.07Chalcedony :−0.008

Siderite :−0.17

Calcite and chalcedony (a quartz polymorph) are also close to equilibrium which is commonlyobserved at low temperatures in this type of rock. The apparent redox potential of the coupleNO3

−/N2(aq) is around 650 mV : significantly far from the specified Eh valueand hence fromredox equilibrium.

56

Chapitre 6. Sorption and ion exchange

We have seen how straightforward it is to include colloidal and mineral solids, how they dissolveand how they can be created by precipitation. Solids have a surface which is in contact with thepore water. Very often, these surfaces are reactive with respect to metals or ligands present inearth’s waters.

Interface reactions between a solute and an inorganic (or organic) surface are of major impor-tance in geochemical systems. Metal sorption by mineral surfaces may significantly reduce themigration velocity of certain species. On the other hand, sorption onto mobile colloids may en-hance migration velocities of contaminants. Both reactions increase the apparent solubility of themetals.

Initially, CHESS has been developed for reactions with colloidal surfaces. The methodologyhas been extended to all types of surfaces, colloidal, mineral, organic or inorganic, and surfacesites or functional groups can be added to all solid species.Very often and especially with clayminerals, thecation exchangeapproach is used to model sorption, or —maybe more preciselysince this process not only includes surface reactions— theretention of metals.

This chapter illustrates how CHESS can be configured for surface complexation and cation ex-change reactions. Also, a short section is devoted to the empirical Kd approach for sorption.

6.1 Surface complexation

Surface complexation is generally considered to be the maininteraction mechanism forpuresystems, such as colloidal (hydr)oxides. Since the publication of the work of Dzombak and Morel(1990) on surface complexation of hydrous ferric oxide, most models use the double layer theoryfor the electrostatic correction term. CHESS makes no exception, applying the formalism to othersolid types as well. In addition, CHESS also provides the constant capacitance and triple layermodel, which can be used if the capacitance values are available. The double layer theory requiresno additional parameters, which can be regarded as an advantage but also as a disadvantage, whenone aims at finding the best fit of experimental data.

The following example concerns a study of nickel sorption onhydrous ferric oxide colloids (50mg/l). A background electrolyte of 5 millimolar NaClO4 is used to maintain the ionic strengthconstant : ClO4− is preferred instead of Cl− in order to avoid Ni-Cl complexes :

Hydrous ferric oxide is thermodynamically unstable with respect to e.g. hematite or other ferric

57


phases as illustrated previously, but the colloids are expected to be kinetically stable given theexperimental time scale. Therefore, dissolution has beenturned offin theOptions panel. Nowrun CHESS and consult the general report. Part of the output file is given here :

Aqueous species:-------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---ClO4[-] 0.0051151 0.005 0.49725 0.5087Na[+] 0.0051151 0.005 0.11495 0.11759Ni[2+] 6.27e-07 6.129e-07 3.5971e-05 3.6799e-05OH[-] 3.4518e-07 3.3741e-07 5.7384e-06 5.8705e-06H[+] 3.1623e-08 3.0911e-08 3.1155e-08 3.1873e-08NaOH(aq) 2.4039e-10 2.3498e-10 9.3985e-09 9.6149e-09Ni(OH)2(aq) 4.7262e-12 4.6199e-12 4.2828e-10 4.3814e-10Ni(OH)3[-] 1.6317e-15 1.595e-15 1.7499e-13 1.7902e-13Ni2OH[3+] 2.6772e-16 2.617e-16 3.5169e-14 3.5978e-14Ni4(OH)4[4+] 3.2269e-23 3.1543e-23 9.5508e-21 9.7707e-21

Minerals and colloids:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams--->HFO 0.00047863 0.00046786 0.05 0.051151

Surface sites:--------------------- ---molal--- ---mol/l--->HFO(w)-OH 1.4166e-05 1.3848e-05>HFO(w)-O[-] 2.177e-06 2.128e-06>HFO(w)-OH2[+] 2.1119e-06 2.0643e-06>HFO(s)-ONi[+] 3.9601e-07 3.871e-07>HFO(s)-OH 4.7818e-08 4.6742e-08>HFO(s)-O[-] 7.3483e-09 7.183e-09>HFO(s)-OH2[+] 7.1284e-09 6.968e-09

At a pH of 7.5, the predominant surface species of ferric oxide is>HFO(w)-OH. The speciationof the surface is strongly dependent on the pH, has shown in, e.g., van der Lee (1998). Anotherexample of surface complexation, over a pH range this time, is discussed in Section 7.1 of thismanual. Part of the nickel isfixedto the colloidal phase, due to a surface complexation reaction.How does CHESS accounts for surface complexation reactionsin this example ? Indeed, nothingof the input data refers to a surface complexation reaction,we only added a certain amount ofcolloids to the system.

Surface complexation is automatically accounted for when the thermodynamic databasedefinessurface complexation for a solid present in the system. For hydrous ferric oxide (and other col-loidal, organic or mineral species as well), the surface complexation reaction is defined in thedatabase1.

1A detailed explanation of how the database is formatted and which keywords are allowed and known by the modelis provided in Appendix A

58

6.1. Surface complexation

The following lines from the database defines species>HFO with, following the suggestion ofDzombak and Morel (1990), strong and weak sites :

>HFO composition = 3 H2O, -3 H[+], 1 Fe[3+]logK = -2.1377vol.weight = 3113.9 kg/m3radius = 10 nmsite >HFO(s)-OH

exch.cap. = 0.093 umol/m2site >HFO(w)-OH

exch.cap. = 3.745 umol/m2

Note the definition of the sites inside the field of>HFO : the sites form integral part of the colloids.Hence by including>HFO in the input script, the model automatically includes the sites in thesystem using>HFO as their mother species. In other words, the sites are included as new basisspecies. Accordingly, it is quite straightforward to definesurface complexes asderived fromthese basis species. In the case of nickel, there is only one reaction defined in the database :

>HFO(s)-ONi[+] composition = 1 >HFO(s)-OH -1 H[+] 1 Ni[2+]logK = 0.37

The complexation of Ni2+ modifies the surface charge of the colloids. As soon as the database in-cludes reactions which allow a change of the surface charge,CHESS will include an electrostaticcorrection (based on the double layer theory). This behavior is easily changed in theSorptionpanel, where the electrostatic correction may be defined by another model or even switched off.

CHESS simultaneously accounts for different types of surfaces, including an unlimited numberof different sites. For example, organic colloids such as Aldrich Humics Acids, denoted in thedatabase by>AHA, can be introduced in the script of the previous example, leading to competitioneffects between organic and inorganic colloids for Ni2+.

59


6.2 Cation exchange

Cation exchange is different from surface complexation : here, a cation from the solid (not ne-cessarily from its surface) isexchangedagainst a cation from the bulk solution. Cation exchangeis often considered to be the predominant interaction mechanism between solutes and clay mine-rals.

The particularity of a cation exchange reaction is the rule of charge conservation : the overall,net charge does not change with the reaction. Mathematically, a cation exchange process canbe compared to a surface complexation reaction without a netchange of surface charge and,consequently, without electrostatics effects. From a physico-chemical view-point, however, thetwo mechanisms are significantly different (see, e.g., Appelo and Postma (1993) for a detaileddiscussion on this subject).

Argillaceous rocks are considered as potential host media for underground repository of radioac-tive wastes. These rocks often contain an important amount of the clay mineral illite, in pure orinter-stratified phases. The evaluation of the confinement properties of the illite fraction with res-pect to a radionuclide such as cesium is an issue of interest in safety assessment : it is importantto know how much cesium is retained by the illite fraction.

First, we must define the main solution in terms of the major species. The porewater is of typeNa-Cl-SO4. The total concentration of Cs+ is estimated at 10µmol/l. Illite is set to a realisticcontent (200 grams ’exposed’ per liter of pore water) and dissolution is not expected to occur(dissolution is disabled in theSolids panel).

The actual cation exchange reaction is implicitly accounted for, via the database —as with thesurface complexation reactions, CHESS assumes that cationexchange data are intrinsic proper-ties of a mineral. However, the actual cation exchange parameters may significantly differ fromone illite to another as a function of the detailed mineralogy of the clay mineral and the rocktype to which it belongs to. Typically, the cation exchange capacity of illite varies from 20 to 50meq/100g in the literature.

It is quite straight-forward to modify the database and introduce our own,ad hocdata. Moreprecisely, we have the two options to do this :

• copy CHESS’ default database to another file, e.g.,Thermo.tdb, and apply all the requiredmodifications to this file. Once the editing finished (you can use the build-in editor ofJCHESS or use your favorite editor instead, i.e. emacs, vi, wordpad, . . .), load the databasevia theChange... button of theDatabase panel. To change the database loaded by default(chess.tdb), add option-d Thermo.tdb to the command-line when launching JCHESS2.

• (re)define species from within the JCHESS environment. Thissimple and save approachis available since JCHESS 2.0 and outlined in the following.

2Windows users may add command-line arguments by editing theshortcut options, Mac users should address the fileoptions.

60

6.2. Cation exchange

To (re)define species, switch to theDatabase panel and open the specific species-define editor byclicking on theAdd... button (or by right-clicking in the table area). The editor allows to definean entire new species or to select an existing species and modify one or several properties. Thisis what we want in this particular example. Hence we click onFind... and select mineral illite.After accepting the selection (Accept), the editor should look at something like the left figure.Several properties of illite are listed. The lower part of the panel displays several panels whichcan be accessed by the tabs. The first panel shows the molal composition of the mineral, togetherwith the log(K), often in polynomial format (log(K) as a function of temperature). Other tabsshow the sites of the mineral, an eventual comment with regard to the species and a reference.Note that all these properties are optional.

For the present example, illite will be extended witha cation exchange site, only one type of site (a sim-plification, because the mineral also has surfacecomplexation sites at the edges which become im-portant at high pH). Note that the cation exchangecapacity is expressed inµmol/m2. The number ofcation exchange sites actually available in the sys-tem is calculated as a function of the specific sur-face area and the total amount of illite present inthe system.

SelectSites and enter the name of the site, e.g.,Illite(Na), with an exchange capacity of 15µmol/m2.To enter the site, double-click on the first emptycell to enterIllite(Na), hit the tabulation keyand type15, hit another time the tabulation key andenter the unit,umol/m2. Hit Enter to validate theresults. The result should look at something likethis :

Validate the modification (click onAccept) whichwill add Illite to the list of (re)defied species in theDatabase panel. At this stage, the illite containsa site, but we have not yet entered any possiblereaction with that site. The actual cation exchangereaction is defined by defining a new site which ef-fectively reacting withillite(Na). Supposing anexchange reaction of Cs+ and Na+, theformation

61


reactionis written as follows :

Illite(Na)+Cs+ Illite(Cs)+Na+

Note that the surface charge remains unchanged with the reaction. A log(K) of 2 is assumed,based on literature data.

This reaction is readily included in the system.Note that in the previous example, we modi-fied an already existing species, while here wedefine an entirely new species. Select surface-sites as theQuantity parameter of the editorand enter the name of the site,Illite(Cs).All other properties, which are optional, areleft empty or to their default values, exceptfor the composing species and logK value. Thecomposition states that 1 mole of Illite(Cs) isformed of 1 mole of Illite(Na) (the site definedas the basis-species of illite), 1 mole of speciesCs+ and−1 mole of Na+. The figure shownhere illustrates the editor after completion. Va-lidate the new surface species (Accept) whichaddsIllite(Cs) to the table of modified or

newly defined species.

At this state, CHESS includes all the information needed to calculate the equilibrium state of acesium-containing solution in presence of illite. As soon as reactive minerals are introduced, onemust also set the available surface, i.e., thevolumetric surfacearea or the surface per liter of porewater. This depends on many parameters, such as packing, porosity and particle size. CHESSallows to set the specific (or volumetric) surface via the concentration editor. Here is how weactually defined the mineral :

Setting the surface to 20 m2/g leads to a volumetric surface area of 200×20 = 4000 m2 per literof solution : quite a lot actually ! With a site-density (or :exchange capacity) with respect to Cs+

of 15µmol/m2, we obtain a total siteconcentrationof 60 mmol/l. The pH is set and several otherspecies are added, leading to the following composition :

62

6.3. TheKd approach

Note that dissolution of Illite is disabled to prevent any phase transformation, not expected tooccur under the current conditions. We verify these calculations with the model results. RunCHESS and examine the general report. Here is part of the report :

Aqueous species:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---Na[+] 0.20535 0.20073 4.6147 4.7209Cl[-] 0.1606 0.15699 5.5657 5.6938SO4[2-] 0.014444 0.014119 1.3563 1.3875NaSO4[-] 0.0060167 0.0058813 0.70019 0.71631NaCl(aq) 0.0030538 0.0029851 0.17446 0.17847Cs[+] 0.00039081 0.00038202 0.050772 0.051941CsCl(aq) 2.5282e-05 2.4714e-05 0.0041607 0.0042565OH[-] 3.4127e-07 3.3359e-07 5.6736e-06 5.8042e-06H[+] 3.9811e-08 3.8915e-08 3.9222e-08 4.0125e-08HSO4[-] 2.2603e-08 2.2095e-08 2.1448e-06 2.1941e-06NaOH(aq) 6.1587e-09 6.0201e-09 2.4079e-07 2.4633e-07HCl(aq) 1.0175e-09 9.9465e-10 3.6266e-08 3.71e-08H2SO4(aq) 6.6984e-19 6.5476e-19 6.4219e-17 6.5697e-17

Minerals and colloids:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---Illite 0.53296 0.52097 200 204.6

Surface sites:--------------------- ---molal--- ---mol/l---Illite(Na) 0.051567 0.050407Illite(Cs) 0.0098141 0.0095933

At equilibrium, the solution has been depleted with respectto aqueous Cs due to the Cs/Na ex-change reaction. The Na exchangeable concentration is onlyslightly modified by the reactionsince the total Cs concentration is low with respect to the total exchange capacity. Another appli-cation of cation exchange is presented in Section 7.3 of thismanual.

6.3 The Kd approach

The preceding sorption reactions were expressed with respect to one or several specific surfacesites. In this way, the full complexity of the chemical reaction and competition effects is takeninto account. Field data, however, is often too scarce to feed the parameters (such as available

63


surface area or site density). Common field data is often obtained for the rock as a whole —andnot for each specific mineral. In that case, only adistribution coefficientor a Kd value is available,describing the ratio between the free and sorbed fraction ofa species. The empiricalKd approachis still commonly used in the field of contaminant transport and retention modelling.

By convention, theKd coefficient is the distribution coefficient between the immobile and themobile concentration of a species. TheKd does not account for site competition or site saturationeffects. The distribution parameter does not refer to a specific species of an element, such asCdOH+, but to the total aqueous concentration of that element, e.g., Cd2+ + CdOH+ + . . .. Dueto these facts, the approach is not valid for conditions beyond those used for determining theKd

value.

A commonly encountered unit forKd is the liter per kilogram. However, an alternative can beconsidered in terms of a non-dimensionalKd, i.e., the ratio between the immobile and the mo-bile concentration of a species, with both fractions expressed in similar units. Hence the non-dimensionalKd is the usualKd (l/kg) times the solid concentration (kg/l).

In order to illustrate the Kd approach, we use our previous example where we added also 0.1µmol/l of another radionuclide, americium. We suppose that theKd of americium in the clay ismeasured and the non-dimensional value yields 200. Distribution coefficients are readily set intheSorption panel. Simply click on the upper-most button to add a new Kd value for a specificelement, e.g. :

Run CHESS for the solution. The general report now includes anew (virtual) species,KdSite-Amwhich represents the fixed Am(III) :

Aqueous species:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---Na[+] 0.20535 0.20073 4.6147 4.7209Cl[-] 0.1606 0.15699 5.5657 5.6938SO4[2-] 0.014444 0.014119 1.3563 1.3875NaSO4[-] 0.0060167 0.0058813 0.70019 0.71631NaCl(aq) 0.0030538 0.0029851 0.17446 0.17847Cs[+] 0.00039081 0.00038202 0.050772 0.051941CsCl(aq) 2.5282e-05 2.4714e-05 0.0041607 0.0042565OH[-] 3.4127e-07 3.3359e-07 5.6736e-06 5.8042e-06H[+] 5.3481e-08 5.2278e-08 5.2691e-08 5.3904e-08HSO4[-] 2.2603e-08 2.2095e-08 2.1448e-06 2.1941e-06NaOH(aq) 6.1587e-09 6.0201e-09 2.4079e-07 2.4633e-07

64

6.3. TheKd approach

HCl(aq) 1.0175e-09 9.9465e-10 3.6266e-08 3.71e-08AmSO4[+] 1.8469e-10 1.8053e-10 6.1211e-08 6.262e-08AmOH[2+] 1.6853e-10 1.6474e-10 4.2833e-08 4.3819e-08Am[3+] 7.4969e-11 7.3282e-11 1.7808e-08 1.8218e-08Am(OH)2[+] 3.4258e-11 3.3488e-11 9.2765e-09 9.4901e-09Am(SO4)2[-] 2.4194e-11 2.3649e-11 1.029e-08 1.0527e-08AmCl[2+] 2.2327e-11 2.1824e-11 6.0771e-09 6.217e-09Am(OH)3(aq) 1.5822e-15 1.5466e-15 4.5474e-13 4.652e-13H2SO4(aq) 6.6984e-19 6.5476e-19 6.4219e-17 6.5697e-17

Minerals and colloids:--------------------- ---molal--- ---mol/l--- ---g/l---- ---grams---Illite 0.53296 0.52097 200 204.6

Surface sites:--------------------- ---molal--- ---mol/l---Illite(Na) 0.051567 0.050407Illite(Cs) 0.0098141 0.0095933KdSite-Am 1.0179e-07 9.9502e-08

As shown by theSurface sites table, CHESS added a hypothetical site namedKdSite-species,KdSite-Am in our case. This site represent the fixed quantity of the total Am concentration. Asexpected, the ratio between the aqueous and fixed fractions of americium is exactly 200. This isreadily read from the table containing the cumulative molalconcentrations :

Cumulative concentrations (molal):----------------------------------

aqueous mineral colloidal organic fixed gaseousH[+] 7.71e-08 0 0 0 0 0Na[+] 0.2144 0 0 0 0 0Cl[-] 0.1637 0 0 0 0 0SO4[2-] 0.02046 0 0 0 0 0Cs[+] 0.0004161 0 0 0 0.009814 0Mg[2+] 4.501e-298 0.1332 0 0 0 0K[+] 0 0.3198 0 0 0 0Al[3+] 1e-25 1.226 0 0 0 0SiO2(aq) 0 1.865 0 0 0 0Am[3+] 5.09e-10 0 0 0 1.018e-07 0

Indeed, we verify that, for Am3+, Kd = 1.018×10−7/5.09×10−10 = 200, exactly. Note that theaqueous fraction of americium has a rather complex chemistry. The related effect on sorption andsolubility is not properly modelled by a simple distribution coefficient.

It is best to avoid the use of distribution coefficients in geochemical models, since theKd is nota thermodynamic quantity. The combined effect ofKd and precipitation/dissolution of specieswhich involve the metal (e.g., Am(OH)3 in our example) is rigorously treated by CHESS, butmay create surprisingly results. Under some circumstances, however, the approach can be useful,

65


especially for reaction path calculations in systems for which distribution coefficients are the onlyavailable sorption parameters.

66

Chapitre 7. Reaction paths

The previous chapters have shown how to apply CHESS to a single equilibrium calculation.More often, we are interested in how aqueous species, minerals or surfaces evolve when one ormore system variables are subject to change. Speciation diagrams, for example, illustrate whichtype of species is predominant for given Eh or pH ranges. In a general sense, areaction pathis asequence of equilibrium states where one or several parameters are modified gradually.

Most of geochemical tools provide specific options to bring pH, species concentrations or thetemperature from an initial value to an end value. CHESS makes no exception, and even extendsthis option : variation of different variables, mixing of two solutions and flushing. Reaction pathmodelling with CHESS is the topic of the present chapter.

7.1 Titration and speciation diagrams

One of the most common reaction paths is atitration, i.e. adding more and more of some che-mical recipe to the solution. For example, a most common titration case is an acid-base titration.With CHESS, it is quite easy to simulate an acid-base titration, i.e., progressively adding HCl orNaOH. However, the titration experiment is generally displayed not as a function of added acidor base, but as a function of pH. CHESS also allows to perform a’titration’ using directly pH asthe dependent variable. We will illustrate this and relatedoptions with help of several examples.

7.1.1 Verification of chemical analysis

Inorganic carbon plays a central role in natural waters, hence the exact value of its concentra-tion is crucial. Unfortunately, a precise measurement of the in situ pCO2 is often difficult, sinceCO2(g) escapes as soon as the sample is brought from the samplingdepth (e.g., several tens ofmeters) to the surface. Analysis shows that inorganic carbonatehasescaped, since the sampleappears to be oversaturated with respect to e.g., calcite ordolomite. We do not know, however,how much exactly escaped from the sample.

CHESS can be used to reconstitute thein situ conditions, as shown in the following example.The result of the chemical analysis is given in Table 7.1. Theelectroneutrality of the solutionrepresents a rough verification of the analysis. We therefore have to translate all concentrations inequivalent units (equivalent = moles× charge), which is tedious and error prone if many speciesare involved. Fortunately, CHESS provides a quick and precise value of the electroneutrality ofthe solution. Enter the chemical analysis in CHESS (Main solution) and run CHESS. Here isa snippet of the general report :

67


pH : 8.22Ca : 42.5 mg/lMg : 3.2 mg/lNa : 13.7 mg/lK : 1.18 mg/lCl : 31.2 mg/lSO4 : 39 mg/lHCO3 : 79.9 mg/lNO3 : 1.3 mg/l

TAB . 7.1– Chemical analysis of a natural water sample from a karstic aquifer.

pH : 8.22ionic strength : 0.00450903temperature : 25 Celsiuselectrical imbalance : -4.6287e-08 eq/lcarbonate alkalinity : 0.0012773 eq/l

The ’goodness’ of the analysis is reflected by the electricalbalance : the value of−2.59×10−5

should be compared with the sum of the equivalents1 (or, approximatively, the ionic strengthof the solution) in order to obtain an estimation of the goodness. For this case, the laboratoryanalysis is very precise : less than 1 % of the solution is unbalanced. The balance could befurther improved by setting a ’balance’ species, e.g., SO4

2−.

Precisedoes not always mean geochemicallyaccurate! Artifacts during analysis or samplingmay have played a role. Indeed, with precipitation and dissolution disabled, we notice that thesolution is over-saturated with respect to several minerals :

mineral Dolomite 0.9412mineral Dolomite-ord 0.9412mineral Calcite 0.3356mineral Aragonite 0.1912

We suppose that inorganic carbon escaped from the sample,beforeanalysis. In other words, thetotal carbonate alkalinity might have been higher than the 0.0012779 eq/l, corresponding to ahigher pCO2. The question is : what is the pCO2 when the solution is deprived of any over-saturated minerals ? A titration with the pCO2 (or : CO2(g)) as variable is a very useful approachto find an answer to this question.

First, we remove HCO3 from the main solution and we impose the atmospheric value ofCO2(g),∼3×10−4 instead. Select theReactions panel, selecttitrate (which is the default reactionpath) and add CO2(g) to the titration panel with an end-value of 0.01.

CHESS requires an output selection, as outlined in chapter 2. A useful selection in this case is thefugacity of CO2(g), pH and saturation indices of e.g. calcite and dolomite.Figure 7.1 illustrates

1Cz[eq] = Cz[mol] × |z|, wherez is the charge of the ion (C).

68

7.1. Titration and speciation diagrams

FIG . 7.1– Reaction panel configured for a titration with CO2(g) as the dependent va-riable.

the final configuration of the reaction panel. Note that CHESSwill liberate the pH, althoughyou imposed its value in the main solution, such that pH will vary as soon as JCHESS starts totitrate the solution with CO2(g). This is the behavior sought for, but opposed to the behavior ofprevious JCHESS versions. It is useless toearly-liberatethe pH in this case, since precipitationand dissolution reactions are disabled.

Run CHESS, which will calculate the initial equilibrium state with a pH of 8.22 and a CO2(g) of3×10−4, followed by a sequence of equilibrium states where, at eachstep, the CO2(g) fugacityis increased to reach a value of 0.01 after 100 steps. The results are written to an intermediate file(CHESS.resby default, you can visualize the file in theOutput panel) but JCHESS provides aplotting package, JPlot, to visualize the results graphically.

Figures 7.2 and 7.3 illustrate the saturation indices and pHas a function of the CO2(g) fugacity.The solution is approximately undersaturated with respectto calcite and dolomite for a CO2(g)fugacity of∼5×10−4±10−4. This threshold value corresponds to a pH of∼8.0. We thereforeconclude that thein situconditions where slightly different from the actual measurements, i.e., ahigher pCO2 pressure and, as a consequence, a lower pH. Hence CHESS can beused to extendour knowledge of natural systems and to detect sampling artifacts.

69


FIG . 7.2– Mineral saturation indices as a function of the CO2 fugacity.

FIG . 7.3– pH as a function of the CO2 fugacity.

70


FIG . 7.4– Reaction path configuration for an acid titration.

7.1.2 Carbonate buffer

The following example uses the titration option to study thebuffer capacity of a carbonate contai-ning solution. A water sample has an initial pH of 10.9 and contains 10 mmol/l of inorganiccarbon and a background electrolyte of 10 mmol/l NaCl. A titration with HCl is used to studythe pH behavior during the acidification process and the buffer capacity of the system.

A main solution is defined by a total CO32− concentration2 of 10 mmol/l, 20 mmol/l of Na+

and 10 mmol/l of Cl− : 10 mmol/l of Na is added to charge-balanced the solution. The initialpH is set to 10.9, but we allow the pH to float during the reaction path (free pH in theOptionspanel). The actual titration reaction is defined in theReaction panel. Selecttitration andadd species HCl to the reaction with an end-concentration of25 mmol/l. The output selection isthe sample number, pH, CO2(aq), HCO3

− and CO32−, in mmol/l. Figure 7.4 illustrates the actual

configuration.

Run CHESS and study the results graphically with JPlot. First, we select pH as a function ofthe sample number. Each sample correspond to 1/100th of the total HCl added to the solutionhence sample 100 corresponds to 25 mmol/l of HCl. For example, Figure 7.5 shows the typicalpH behavior of a carbonate-buffered solution.

2This is not the actual concentration of species CO32−, see Chapter 2, section 2.1 for a discussion on the difference

betweentotal concentrationandconcentration.

71


FIG . 7.5– pH behavior during a titration of carbonated water with HCl. 100 samplescorrespond to 25 mmol/l of added HCl.

FIG . 7.6– Speciation of the carbonate system as a function of pH.

72


Visualization of the selected carbonate species as a function of the added acid or, as illustratedby Figure 7.6, as a function of pH is a straightforward task with JPlot. Figure 7.6 shows the spe-ciation diagram for the inorganic carbon species. You easily recognized the classical distributionof the carbonate species as pH shifts from alkaline to more acidic conditions.

7.1.3 Uranium speciation

A speciation diagram can be obtained for any system as a function of many different variables.For example, the complexation of uranyl ions by carbonate ligands can be studied as a function oftotal carbon concentration. Here we take a simple solution containing 1µmol/l of total UO2

2+,a pH of 8 and an initial total HCO3− concentration of 10−6 mol/l. This example shows twointeresting new features :

• the logarithmic option, useful to increase the graphical resolution for logarithmic scalingbut also to help CHESS to achieve it’s work for very reactive systems. With the logarithmicoption enabled, CHESS brings the total carbonate concentration from 10−6 to 10−1 mol/lwith a logarithmically increasing step size.

• the wild-character “*” can be used to select all the species including the substring ”UO2” intheir name, thus including all species such asUO2[2+], UO2CO3(aq), etc.. Such a selectionmust be introduced manually : selectaqueous-species, erase the item name, type*UO2*and select an appropriate unit :

Figure 7.7 shows this specific diagram, obtained for 1µmol/l of total UO22+ and an end-value of

the total carbonate concentration of 10−1 mol/l. The competition effect of carbonate complexa-tion with respect to hydroxyl complexation is enhanced as the carbonates aqueous concentrationincreases, leading to the predominance of UO2(CO3)3

4− which stabilizes uranium and favors itstransport in groundwater.

7.1.4 Surface complexation as a function of pH

A speciation diagram may also include one or several solid species, possessing reactive surfacesites. For example, let us consider the evolution of the fraction of copper and selenium sorbed tohydrous ferric oxide colloids as a function of pH. The main solution is defined as follows :

73


FIG . 7.7– Speciation diagram of the complexation of uranyl ions by carbonate li-gands.

pH = 4total Na+ = 10 mmol/ltotal Cl− = 10 mmol/ltotal SeO4

2− = 5 µmol/ltotal Cu2+ = 5 µmol/lcolloid HFO = 1 g/l, surface = 200 m2/g

Furthermore, precipitation is disabled and the system is balanced via species Na+. The chargebalancing feature remains enabled during the titration. This feature is very useful for titrationwith e.g. pH, because increasing the pH creates a charge deficit (since hydroxyl ions are addedto the solution).

The reaction path is a base-titration but instead of adding e.g., NaOH, we directly use pH whichis progressively increased to 8.5. The output selection is pH, thesurface chargeof hydrous ferricoxide inµC/cm2 and the colloidal bound fractions of SeO4

2− and Cu2+ in µmol/l (fixed fractionsof these species). These items are readily selected with help of theOutput selector.

Hydrous ferric oxide is an amphoteric colloid, positively charged at low pH and negatively char-ged at hight pH. Consequently, and illustrated by Figure 7.8, the positively charged Cu2+ ion iscomplexed at high pH while the anionic SeO4

2− strongly complexes at low pH. We recall thatprecipitation and dissolution has been disabled : without this option, HFO would dissolve to formhematite, thus destroying all reactive surface site.

74


FIG . 7.8– Evolution with respect to pH : a) electrical surface chargeof HFO colloids,b) fixed fractions of Cu and Se.

75


The CHESS database includes many surface complexes for hydrous ferric oxide and for someother colloids, such as amorphous silica, quartz, clay particles and organic matter.

7.2 Polythermal reaction paths

The geochemical evolution of aquatic or mineral systems under thermal stress is often referred toas polythermal reaction paths. Temperature is an intensivevariable : many thermodynamic (andrelated) processes in aquifer systems are temperature dependent. CHESS accounts for depen-dence of log(K)’s as soon as some kind of polynomial relationship is provided by the database.To find out whether a species is defined as a function of temperature, simply check the databaseusing theAnalyzer. Configuring a reaction path with the temperature as the dependent variableis similar to the approach outlined in the former sections.

Configuring a polythermal reaction path with CHESS isillustrated at the hand of the following example, whichaims at understanding one of the daily problems annoyinghumanity : calcification of the shower-head. Tap water(our main solution) is assumed to originate from limes-tone aquifers near Fontainebleau, has a pH of∼7.3 andcontains 92 mg/l of calcium, 260 mg/l HCO3. Our star-ting (aquifer) temperature is 20C. We further simplifythe example by excluding all solids but Calcite. This rea-

dily done in theDatabase panel, as shown by the screen shot besides. Thusly, the actual databaseis drastically reduced, a very useful option in many situations.

The polythermal reaction is defined in theReactions panel. Select the temperature with an end-value of 80 C. The output selection is the temperature, the concentration of H+ and mineralcalcite, both in mol/l. Now run CHESS and examine the resultsusingJPlot (see Figure 7.9).This graph shows that, as the temperature increases, the saturation indices of calcite increaseleading to calcite precipitation. Thisretrogradebehavior of a mineral solubility with temperature—surprising at first sight— is due to the indirect effect of CO2(g) degazing.

Other factors, such as precipitation reaction kinetics, also play a role in this process which poi-sons your shower-pleasure. But before we recommend to take short showers only using a watertemperature below 10C, we suggest to perform these calculations using your localtap-waterchemistry : the outcome can be significantly different.

7.3 Flushing

Flushing is a process where, at each step, part or all of the mother solution is removed andreplaced by the solution defined as theflushor secondarysolution. As outlined in Chapter 2, thevolume ratio of the main- and flush solutions is of great relevance : for each sample, a fraction—or the total volume— of the mother solution is replaced by the flush solution. The size of the

76

7.3. Flushing

FIG . 7.9– Precipitation of calcite as a function of temperature.

fraction is defined by the ratio of the main- and flush solutions and the number of samples. Letvbe the volume fraction replaced at each step, then

v =flush solution volume

(main solution volume)×N(7.1)

whereN is the number of samples. Hence, if the main solution volume is 1 liter, the flushingvolume 10 liter andN is 10, the entire solution is replaced at each step. It is not possible toreplacemorethan the main solution, i.e. CHESS imposes thatv≤ 1. After each flushing step, theequilibrium state of the resulting solution is calculated,assuming total and immediate mixing.

For example, let us consider the pollution of a fresh water aquifer by salt water (see for instanceAppelo and Postma (1993)). As a result of calcite dissolution, the aquifer solution is characterizedby Ca2+ and HCO3

−. The solid, clayey aquifer material acts as a cation exchanger and, giventhe water type, most sites are occupied by Ca2+.

In sea water, however, the major ions are Na+ and Cl−. When sea water intrudes the aquifer, anexchange reaction of cations takes place, which can be written in terms of the Gaines-Thomasconvention as follows :

Na+ +0.5 X(Ca) X(Na)+0.5Ca2+. (7.2)

We are interested in modelling this process of water intrusion and cation exchange in terms offixed and aqueous cation fractions. First we define the main solution, i.e., the aquifer, as follows :

77


pH = 8total Na+ = 10 mmol/lmineral Quartz = 800 g/lmineral Calcite = 150 g/lmineral Clay = 50 g/l, surface = 15 m2/g

Cation exchange reactions are characterized with respect to an average bulk material —correspondingto the global response of the clay and organics fractions of the soil or rock—, rather than withrespect to specific mineral phases. In our example, we have introduced a generic clay mineral,namedClay, defined in the database with an appropriate cation exchangesite. The exchange ca-pacity of Clay(Ca) has been set to 10µmol/m2, which corresponds to 20µeq/m2 or 20×15= 300µeq/g. MineralClay and the secondary surfaced siteClay(Na) do not exist in the database, theyare nevertheless readily introduced in the system using the(re)defineoption of theDatabasepanel :

The actual reaction path, with the sea-water type flushing solution, is defined in theReactionspanel. The solution is composed of 480 mmol/l Na, 10 mmol/l Caand 500 mmol/l Cl. The pHis 7 and the flushing volume is 5 liters. Figure 7.10 illustrates the configuration of the reactionpath, as well as the output selection.

Note that, in CHESS, the activities of the fixed ions are calculated with respect to the usualstandard state of 1 mol/kg of H2O, and not with respect to the solid like in the Gaines-Thomasconvention (van der Lee 1998). In the case of Eq. 7.2, the transcription of a Gaines-Thomasselectivity coefficient,KGT, into a cation exchange constant,K, is

K = KGT√

2σAsS.

whereσ denotes the exchange capacity of the clay (only of site,Clay(Ca)), As is the specificsurface andS is the total clay concentration.

78

7.3. Flushing

FIG . 7.10– Configuration of the reaction path for a salt water intrusion in a clayeyfresh-water soil.

The effect of salt water intrusion is illustrated in Figures7.11 and 7.12. The bicarbonate calciumwater shifts towards a Na-Ca-Cl water type. The incoming sodium cations replace the exchan-geable calcium ions, which now enter the solution. Nevertheless, the total aqueous concentrationof Ca2+ remains controlled by mineral calcite during the flushing process.

Flushing is only a first step towards the simulation of the true hydrogeological system dynamics.Accurate modelling of intrusions or pollution dissemination requires a transport model coupledwith geochemistry, as outlined in e.g., van der Lee (1998).

79


FIG . 7.11– Evolution of aqueous concentrations of sodium and calciumcations duringsalt-water intrusion in a fresh-water aquifer, modelled bya flushing reactionpath.

FIG . 7.12– Evolution of fixed concentrations of sodium and calcium cations.

80

7.4. Mixing of two solutions

7.4 Mixing of two solutions

Mixing is a process where, at each step, the main solution ismixedwith a fraction of the solutiondefined as themix or secondarysolution. At the end of the reaction (last sample), both solutionsare entirely mixed.

CHESS first solves the equilibrium state of the main solution, then for secondary. The secondarysolution is (linearly) added to the main solution inN steps whereN is the number of samples(100 by default). This option is especially useful for complex geochemical systems, where twosignificantly different waters are mixed (for simple systems, a titration-like reaction path can beused). It is interesting to see that the final solution is not very often equivalent to a conservativemixture of two solutions, especially when precipitation, dissolution or adsorption processes areinvolved.

By default, CHESS mixes both solutions entirely, includingmineral, gaseous and colloidal phases.Sometimes, it is useful to mix only themobilefraction of the secondary system. The mobile frac-tion includes only the species which are ”displaceable” by aprocess such as hydrological flowor diffusion. This concerns all species except for mineralsand species sorbed to these minerals.Note that colloids form part of the mobile fraction. In orderto mix with mobile fractions it isnecessary to checkmobile-fraction.

For example, a society wants to predict the solubility of barium in a limestone aquifer afterhaving been polluted by a mine waste solution. The wastes originate from an acidic treatmentof ore deposits and the corresponding seepage water is strongly enriched in barium. The mixingratio of 1 :9, i.e., one liter of mine waste solution is mixed with 9 liter of aquifer solution.

First, we define the main solution which is simply a calcite-type water (e.g.,Calcite = 1 kg/l)with a pH of 7.9, a temperature of 10C. Furthermore, we set the volume of this solution to 9liter.

Note that the pH of the solution remains fixed at 7.9, also during the flushing process. This is notwhat we want : you must set the pH to float in the Options panel (free pH).

The secondary solution, used for mixing, is defined in theReactions panel. Click onmixing,which opens a panel similar to theMain solution panel. Here we enter a barium-containingsolution, at 18C and a pH of 2. The secondary solution volume is kept at 1 liter: accordingly,we will have exactly 10 l of solution after complete mixing. We select the number of samples(each sample corresponds to a degree of mixing), the pH and the aqueous as well as the mineralfraction of barium-containing species. Note that the temperature of the main solution is calculatedduring the mixing process by CHESS, assuming identical heatcapacities for each fluid.

The results are illustrated Figure 7.13, showing a progressive decrease of pH as the mixing pro-cess proceeds. The final pH differs from the value obtained after a conservative calculation,which is due to the buffering role of calcite. Table 7.2 provides the mineral and aqueous frac-tions before- and after mixing of the solutions. The barium solubility is controlled by the mineralwitherite, which is a barium-carbonate mineral, only slightly soluble in neutral pH conditions.The barium concentration in the aquifer is thousand times lower than the polluted water. Notethat the aqueous barium concentration does not remain constant during the mixing process due

81


FIG . 7.13– Evolution of pH during the intrusion (mixing) of a barium-containing solu-tion in a karstic aquifer.

Wastes Aquifer Mixtureaqueous fraction (mg/l) 2.5 0. 0.002

∑mineral fraction (mg/l) 0. 0. 0.248

TAB . 7.2– Aqueous and mineral fractions of barium before and after the mixing pro-cess.

to a decrease in CO32− with pH.

What if we want to mix the resulting solution with another water type ? Well, the bad news isthat the current version of JCHESS only handles two solutiontypes. The good news, however, isthat JCHESS provides a general option which allows to ’pick up’ the resulting solution (henceafter mixing in our example) using theresumeoption. This option creates a new and valid inputscript, namedresume.chsunless you set another name. This file can be loaded in JCHESS as anew main solution.

82

Chapitre 8. Kinetics of dissolution and precipitation

So far, all the simulations carried out by CHESS were based ona thermodynamic equilibriumapproach. It is well recognized, however, that many reactions arekinetically controlledover thesimulated time range. CHESS deals with kinetics of dissolution and precipitation reactions forboth mineral and colloidal species.

8.1 Kinetic laws

The mechanisms involved in kinetic dissolution or precipitation of a solid are generally complex.For instance, the precipitation of a mineral depends on the nucleation process and the saturationstate of the solution with respect to the mineral. The saturation state for a mineralMpXq, isdefined by the Ion Activity Product (IAP) :

Ω =IAPKs

=[Mq+]p[Xp−]q

Ks, (8.1)

where[Mq+] is the activity of the ionMq+ andKs the solubility constant. Note thatKs is 1/K, Kbeing the formation constant used throughout this documentand in CHESS’ databases.

The saturation state is also used as an indicator of the tendency for a mineral to dissolve. When thesolution is under-saturated with respect to a mineral, the mineral tends to dissolve. The saturationstate is thus a variable which should be taken into account. The following general kinetic reactionfor a solidS(mineral, colloidal) is used by the code :

dSdt

=

ApkpWp(Ωa−1)b if Ω ≥ 1

−AdkdWd(1−Ω f )g if Ω < 1(8.2)

where subscriptsp and d refer to precipitation and dissolution, respectively. Thekinetic rateconstant is denoted byk in mol/m2/s, A is the volumetric surface area expressed in m2/m3, a,b f andg are arbitrary power-constants, used to fit the law to experimental data,Ω is the satu-ration state as mentioned above.W introduces an eventual dependence on specific ion activitiesfrom solution. It actually stands for∏i [Ci ]

ai j and represents thesolution-dependingfactor of theapparent reaction rate. For a positive power (ai j ) of species (Ci), the species will catalyze thereaction rate. For a negative power, the species inhibits the reaction rate. For example, a precipi-tation reaction catalyzed by the presence of O2(aq) and protons could contain aW term definedas follows :

Wp = [O2(aq)]0.65[H+]0.4

By default, all the power-constants as well asW are set to 1. Note that, for ab or g of 0, the lawbecomes entirely independent of the saturation state.

As illustrated by equation (8.2), the precipitation law is not necessarily the same as the dissolutionlaw, even if it concerns one single solid phase. This allows to simulate mineral reactions with fastprecipitation, but slow dissolution kinetics, for example.

83


FIG . 8.1– Configuration of the kinetics editor for calcite dissolution.

8.2 Calcite dissolution

The impact of reaction kinetics is illustrated with help of asimple example : calcite dissolutionunder controlled conditions. The aim of the study is to quantify the calcite dissolution behaviorin open (buffered) and closed (unbuffered) systems. We are also interested in the temperaturedependence of this particular system. The background solution contains 40 mmol/l NaCl, isequilibrated with the atmospheric pCO2 (i.e., a CO2(g) fugacity of∼ 3×10−4) and a pH of7.5. Thus specified, the pCO2 and pH are kept constant (buffered) during the reaction. 100gramsof calcite is added to 1 liter of this solution, its specific surface is set to 50 cm2/g, correspondingto millimeter-sized grains.

Calcite dissolution is kinetically controlled. The kinetic control of a solid species is specified inthe lower part of theSolids panel. Click on ’Add...’ (or right-click within the table area) toopen theKinetics editor. Select mineral Calcite from the list. Only one type of reaction lawis currently available with JCHESS, the default law. The third field indicates whether the lawis valid for precipitationand dissolution, precipitation-only or dissolution-only. Inthe presentcase we use the default option, knowing that system does not lead to precipitation anyway. Thekinetic rate of calcite dissolution is set to 1.4×10−9 mol/cm2/s. We also set the power ofΩ,f in the list of dependent species : typeomega instead of a species and set its value to 0.51.Figure 8.1 illustrates the configuration of the kinetic law for calcite. Click onApply to validatethe parameter configuration, which adds calcite to a list of kinetically controlled species.

In order to study the dissolution behavior, we perform a reaction path withtimeas the dependent

1For a more detailed explanation of the different parameters, the reader is referred to section 2.2.

84

8.3. Plagioclase albitization

variable. This reaction is included in CHESSas if it was a titration reaction. Hence, in theReactions panel, click ontitrate and select itemtime from thevariables context. Notethat the start value of a reaction with time isalways02. Select an end-value of 20 minutes. Theoutput selection is time in minutes, the aqueous fraction calcium and the saturation index ofcalcite as illustrated here :

The results of the simulation are displayed in Figures 8.2 and 8.3. In buffered conditions, the Ca-release rate remains approximately constant during the reaction. In a closed, unbuffered system,however, the reaction stops (no dissolution) after a few minutes, due to the rapid increase of pHwhich reaches a value close to 10. Accordingly, the solutionapproaches equilibrium with respectto calcite (Ω = 1, a saturation index of 0) which stops the dissolution reaction.

8.3 Plagioclase albitization

The following example shows another application of precipitation- and dissolution kinetics. Wepropose to study the geochemical processes of sandstone diagenesis in contact with saline fluidsfor a medium temperature range. The transformation reaction is written as follows :

Anorthite+2Quartz+0.5H2O+Na+ +H+ → Albite+0.5Kaolinite+Ca2+,

which is also known as the process of plagioclase albitization.

Here we are interested in the kinetically controlled minerals only, hence we propose to simplifythe system considerably : exclusion of all minerals and colloids (added to the exclude list of theDatabase panel), except for Albite, Anorthite, Kaolinite and Quartz(added to the include list).

The main solution, added to theMain solution panel of JCHESS, is resumed in table 8.1.Note that theconcentrationof mineral kaolinite is imposed : this means that we fix the actual

2It is nevertheless possible to set a pre-equilibration timeof the main solution using the ’time’ variable in the main-solution panel.

85


FIG . 8.2– Aqueous Ca during calcite dissolution under buffered (pH and pCO2) andunbuffered conditions.

FIG . 8.3– Saturation index of calcite under buffered and unbufferedconditions.

86

8.3. Plagioclase albitization

pH = 7.8temperature = 80 C

total Na+ = 530 mmol/ltotal Cl− = 560 mmol/ltotal Ca2+ = 13 mmol/lconcentration Kaolinite = 50 g/lmineral Quartz = 750 g/l, surface = 200 cm2/gmineral Albite = 100 g/l, surface = 200 cm2/gmineral Anorthite = 100 g/l, surface = 200 cm2/g

TAB . 8.1– Main solution composition of an example of plagioclase albitization.

concentration of kaolinite, whatever the amount of composing species in solution3. Indeed, theinitial fluid is assumed to be in equilibrium with respect to kaolinite, since the kinetic of kaoliniteis much faster relatively to the other minerals.

Kinetic behavior is attached to minerals quartz, albite andanorthite using the kinetics editor in theSolids. To somewhat simplify the case, all minerals and colloids are excluded from the systemin theDatabase panel. You can safely do this : CHESS will still include all minerals defined ineither the main solution, either the secondary solution or titration reaction. For minerals albite,anorthite and quartz we set a symmetric rate constant (i.e. for precipitationanddissolution) to10−15, 10−15 and 5×10−16 mol/cm2/s, respectively. All other kinetic parameters are not usedorleft to their default value.

The reaction is studied as a function of time : we therefore set up a reaction path with time :a titration-like reaction path using time as the principal variable, as outlined in the previoussection. The time-span of interest is 1500 years, and the output selection is the time in years, pH,saturation indices of the four minerals as well as their concentrations in g/l. All these items arereadily selected with help of theOutput selector.

Run CHESS and study the reactions with the graphical front-end JPlot. Figure 8.4 illustratesthe evolution of the saturation indices and the mineral concentrations as a function of time.Once the kinetic reactions start, anorthite and albite —thesolution is respectively undersaturatedand oversaturated with respect to these minerals— dissolveand precipitate. It is worth noticingthat quartz, which was initially in equilibrium, also dissolves due to the relative differences inthe kinetic rates of each minerals. During the reaction-path, the pH of the solution increasesslightly (the pH evolution is however more complex than in the previous example). At the end ofthe simulation, all the minerals have reached thermodynamic equilibrium. Since the kinetic lawdepends on the surface area of the mineral, a mineral cannot precipitate if not present from thestart. CHESS proposes different solutions to this problem :

• specify a nucleus surface:you can specify the nucleus surface, which is one of the kinetic parameters proposed in thekinetics editor. The nucleus surface is aminimumsurface on which the mineral can

3For a detailed discussion on qualifiers such asconcentration, the reader is referred to, e.g., section 2.1.1.

87


FIG . 8.4– Modelling of kinetically controlled plagioclase albitization at 80C.

88

8.4. Temperature dependence

A = A x S

vs

A = nucleusv

S

surf

ace

surface used

by reaction

FIG . 8.5– Schematic of the use of a nucleus surface in CHESS.Av is the volumetricsurface area (m2/l), As the specific surface area (m2/g) andS is the mineralconcentration.

precipitate. As soon as the mineral start to precipitate, CHESS calculates its actual surfacearea. The kinetic rate law will use the actual surface area assoon as it exceeds the nucleus.Figure 8.5 illustrates, graphically, the way the model usesthe nucleus surface.

• specify a static surface:in an aquifer system, minerals will tend to precipitate on the surface of the solid matrix,

i.e., the grains of the porous medium. Accordingly, the kinetic rate is more or less inde-pendent of the surface of the precipitating mineral itself.This surface can be set in themain solution as follows :

– define the mineral with a zero concentration

– set its surface in m2 per liter of solution.

Accordingly, the surface is fixed and remains independent ofthe amount of precipitatedmineral.

8.4 Temperature dependence

CHESS implements the Arrhenius law to handle temperature dependence of kinetic reactions.This law is written as follows :

k = Aexp

(

−EA

RT

)

, (8.3)

whereA is a pre-exponential factor,EA is the apparent activation energy (in J/mol),R is theperfect gas constant, andT is the absolute temperature inK.

What would be the impact of temperature on the calcite dissolution rate constant ? Accordingto the configuration of the solution described in section 8.2, we have an intrinsic dissolutionrate constant valid for 25C only. Extrapolation to higher temperatures, e.g. 60C, is quitestraightforward in CHESS, only two steps are needed :

• set the temperature for which the rate is actually valid, hence 25C in our case. In absenceof a temperature value in kinetics editor, CHESS assumes that the rate corresponds to thetemperature of the main solution ;

89


FIG . 8.6– Total aqueous concentration in calcium during calcite dissolution at 20 and60 C.

• set the Arrhenius energy parameter (obligatory).

The evolution of the aqueous concentration of element Ca at 20 and 60C is plotted in Figure8.6. The dissolution of minerals generally increases with temperature (they become less stable).Indeed, calcite dissolution at 60C is roughly 10 times faster compared to the process at 20C,at least at the start of the simulation and for buffered pH andpCO2.

8.5 Catalyzing and inhibiting effects

It is often observed that the kinetic rate can be strongly affected by the presence of species insolution, even if they do not form part of the mineral composition. Such species have acatalyzingeffect when they speed up the reaction, or aninhibiting effect when they tend to slow downthe reaction. Catalyzing or inhibiting species are taken into account by CHESS, which uses ageneralization of the basic kinetic law (see equation (8.2)).

We recycle the example of calcite dissolution in order to illustrate the impact of this feature.Calcite dissolution depends also on pH, hence if the pH is variable, this dependence should betaken into account.

Catalyzing or inhibiting species are included in the reaction via theKinetics editor, i.e.,the table entitledDependence. Here we introduce species H+ with a power of 0.5. Figure 8.7illustrates the editor, fully configured for temperature and solution dependence. The results ofthe current model configuration are shown in Figure 8.8 for two pH values (note that the pH isbuffered during the reaction). The dissolution rate of calcite increases significantly with pH, as it

90

8.5. Catalyzing and inhibiting effects

FIG . 8.7– Configuration of the kinetic rate law of calcite with catalyzing species H+

and temperature dependence.

is expected from equation (8.2).

Kinetic constraints on the dissolution or the precipitation of solids can be used together withmixing or flushing reaction paths. This interesting featureof CHESS is particularly useful if youneed to studydynamicsystems, such as transport of contaminants in an aquifer or flow reactors.

The overall duration of the reaction path is set using thetime option in the main solution panel.For example, mixing of two solutions with 100 samples and an overall duration of 100 days leadsto 100 steps where, at each step, the system reacts for a period of exactly 1 day.

91


FIG . 8.8– Aqueous Ca during calcite dissolution for two different pHvalues. De-monstration of reaction catalyzing effects.

92

Appendix A. Thermodynamic databases

A.1 Available databases

Initially, CHESS was provided as a solver only without any database. Accordingly, the responsi-bility of which database should be used was left to the geochemical wisdom of the user. However,this approach appeared to represent a major handicap for novice users, who had to learn not onlyhow to run CHESS but also how to format a database. It was then decided to provide a CHESSdatabase based on the LLNL (EQ3/6) database but enriched with respect to sorption sites (surfacecomplexation, cation exchange) and colloidal species.

Today, several public available databases are available for use with CHESS. They are all availableathttp://chess.ensmp.fr, and only briefly listed here :

• chess.tdbis a CHESS-formatted version of EQ3/6 (V.8-R.6) database (Wolery 1992). Thisdatabase is moderated by disregarding a number of organic redox-species (see Table A.1)and their derived species, and extended with colloids and sorption sites. This databasecontains 1220 aqueous species including 98 redox couples, 1126 minerals, 91 gases, 74sites, 3 inorganic colloids, and 3 organic colloids ;

• eq36.tdb is a CHESS-formatted version of the full EQ3/6 (V.8-R.6) database (Wolery1992) ;

• minteq.tdb is CHESS-formatted version of the MINTEQ database, a specific sub-surfaceoriented database which comes with the MINTEQ software (Allison et al. 1991) ;

• phreeqc.tdbis CHESS-formatted version of the PHREEQC database, the onewhich shipswith PHREEQC ;

• nea.tdb is a CHESS-formatted version from the NEA thermochemical database projectTDB (see http ://www.oecdnea.org).

• wateq4f.tdb is a CHESS-formatted version of the WATEQ4F code.

Other databases are under development and made available tothe CHESS users as soon as theyreach a stable state. Supplementary information on specificdatabases, as well as updates, can befound on on the CHESS web site.

Acetic acid(aq) Acetone(aq) Alanine(aq) Benzene(aq)Butanoicacid(aq) Ethanamine(aq) Ethane(aq) Ethanol(aq)

Ethylene(aq) Ethyne(aq) Formic acid(aq) Glycolic acid(aq)Glycine(aq) Lactic acid(aq) Malonic acid(aq) Methane(aq)

Oxalic acid(aq) Pentanoicacid(aq) Phenol(aq) Propanoicacid(aq)Succinicacid(aq) Toluene(aq) oPhthalate[2-]

TAB . A.1 – Organic species present in the full LLNL EQ3/6 database, but not includedin the moderated CHESS database.

93


H2O Ag[+] Al[3+] Am[3+] Ar(aq)Au[+] B(OH)3(aq) Ba[2+] Be[2+] Br[-]Cd[2+] Ca[2+] Ce[3+] Cl[-] Co[2+]

CrO4[2-] Cs[+] Cu[2+] Dy[3+] Er[3+]Eu[3+] F[-] Fe[2+] Ga[3+] Gd[3+]H[+] H2AsO4[-] HCO3[-] HPO4[2-] He(aq)Hf[4+] Hg[2+] Ho[3+] I[-] In[3+]K[+] Kr(aq) La[3+] Li[+] Lu[3+]Mg[2+] Mn[2+] MoO4[2-] NH3(aq) Na[+]Nd[3+] Ne(aq) Ni[2+] Np[4+] O2(aq)Pb[2+] Pd[2+] Pr[3+] Pu[4+] Ra[2+]Rn(aq) Rb[+] ReO4[-] RuO4[2-] SO4[2-]

Sb(OH)3(aq) Sc[3+] SeO3[2-] SiO2(aq) Sm[3+]Sn[2+] Sr[2+] Tb[3+] TcO4[-] Th[4+]

Ti(OH)4(aq) Tl[+] Tm[3+] UO2[2+] VO[2+]WO4[2-] Xe(aq) Y[3+] Yb[3+] Zn[2+]

Zr(OH)2[2+] O2(aq)

TAB . A.2 – List of the basis species of the full CHESS database

Without any specific needs, the use ofchess.tdbis recommended. Most of the sorption data havebeen taken from Dzombak and Morel (1990), but new data (mainly involving colloids) have beenadded as well. The set of default basis species are conventionally the building blocks of all thesecondary species of the database. Table A.2 lists the default basis species of the full CHESSdatabase.

It is not necessary to select the principal components from the default set of basis species, becauseCHESS knows how to implicitly swap species from the derived field to the basis (i.e.speciespromotion). This is how minerals, colloids and gases are included in the system. Very often, it iseven illogical to select default basis species. Think, for instance, to Fe2+ (written asFe[2+] inCHESS format) in a highly oxidizing environment . . . The model will use valuable time to findout that it is more likely to use hematite or another hydrous ferric oxide instead, something whichcan be anticipated by the user. A choice of principal components driven by chemical tuition issometimes even necessary in order to achieve convergence, especially in redox systems.

As discussed in Chapter 5, the natural waters are often in disequilibrium with respect to some(or many) redox species. In that case, it would be useful todecouplea specific, redox sensitivespecies from the redox equilibrium reactions. Once a redox coupled has been disabled, CHESSconsiders the disabled species as equivalent to a basis species. The disabled species has thus itsown mass balance and independent entry in the input script. The coupled redox species, presentin chess.tdb, are listed Table A.3.

94

A.1.

Available

databases

−III −II −I 0 I II III IV V VI VIIAg[+] Ag[2+]

Am[2+] Am[3+] Am[4+] AmO2[+] AmO2[2+]AsH3(aq) H2AsO3[-] H2AsO4[-]

Au[+] Au[3+]Br[-](1) BrO[-] BrO3[-] BrO4[-]

CO(aq)(2) HCO3[-]Ce[2+] Ce[3+] Ce[4+]

Cl[-] ClO[-] ClO2[-] ClO3[-] ClO4[-]Co[2+] Co[3+]Cr[2+] Cr[3+] CrO4[3-] CrO4[2-]

Cu[+] Cu[2+]Fe[2+] Fe[3+]

H2(aq) H[+]Hg2[2+] Hg[2+]

I[-](1) IO[-] IO3[-] IO4[-]Mn[2+] Mn[3+] MnO4[2-] MnO4[-]

NH3(aq) N2(aq)(1) NO2[-] NO3[-]Np[3+] Np[4+] NpO2[+] NpO2[2+]

H2O(aq) O2(aq)Pb[2+] Pb[4+]

Pu[3+] Pu[4+] PuO2[+] PuO2[2+]REE[2+](3) REE[3+]Ru[2+] Ru[3+] Ru(OH)2[2+] RuO4[2-] RuO4[-]

HS[-] S2[2-](1) S2O3[2-](4) S2O4[2-](1) SO3[2-] S2O6[2-] SO4[2-] S2O8[2-]HSe[-] SeO3[2-] SeO4[2-]

Sn[2+] Sn[4+]Tc[3+] TcO[2+] TcO4[3-] TcO4[2-] TcO4[-]

Tl[+] Tl[3+]U[3+] U[4+] UO2[+] UO2[2+]V[3+] VO[2+] VO2[+](5)

Table A.3 : Full list of the redox species from the CHESS database. The oxidation states of the elements are denoted byRoman numbers. (1) Some redox series have additional species with fractionnal oxidation states, i.e.,Br3[-],I3[-], N3[-], S4[2-], S5[2-], S4O6[2-] andS3O6[2-]. (2) AlsoCN[-] andSCN[-]. (3) REE = Dy, Er, Eu, Gd,Ho, La, Nd, Pm, Pr, Sm, Tb, Tm, Yb. (4) AlsoS5O6[2-]. (5) AlsoVO4[3-]

.

95


A.2 The CHESS database format

CHESS uses a specific, keyword based database format. The keyword-based approach of for-matting the database has a number of important advantages. First, the database is more compact(about one third of the size of the original LLNL version), hence its reading is much faster. Se-cond, the format is evolutive, i.e. new attributes can be added without loosing “downward compa-tibility” with previous databases. The format guarantees an optimal readability and the keywordsare intuitive, hence easy to remember. Updating the database with new data is therefore straight-forward even for novice modelers. And last but not least, thekeyword based approach allows forconversion to other formats for other models.

The data is compiled in ASCII format in a file of over 15,500 lines (about 550 Kb). The file startswith comments to identify the origin and version of the database and a series of general constants.The CHESS database is then structured according to major fields. A field is a specific area delimi-ted by brackets. Everything within brackets belongs to its specifier, which is the name of the field.The following major fields are present :basis-species, redox-couples, aqueous-species,gaseous-species, minerals, colloids, organic-matter andsurface-sites.

Each major field consists of a number of minor fields, which arespecies with the type of the majorfield. For example, the fieldaqueous-species contains the speciesAl(OH)2[+]. All attributesand characteristics of the species are given in its associated field :

Al(OH)2[+] composition = -2 H[+], 1 Al[3+], 2 H2OlogK = -12.1394(0), -10.5945(25), -8.7455(60), -6.9818(100)\

-5.146(150), -3.5858(200), -2.2019(250), -0.9295(300)radius = 2e-10 m

A number of general formatting rules apply to the database. Each major field must be openedwith ”” and closed with ””. Comments are possible using the comment-specifier ”#”. Theorder of the fields is not fixed, you may change it. But the actual order is optimized for readingby CHESS. Empty lines and white spaces at the start of each line are ignored. The special line-break character, ”\”, extends reading to the next line. The logarithms of formation constantscan be introduced forarbitrary temperatures(in C) in the database. CHESS used this list tocompute by interpolation the formation constant for any temperature between the minimum andmaximum temperature provided by the polynomial description. The log(K) list can be reducedto one single value, which will be used for all temperatures.

Here is a list of the minor field structure for each major field :

Basis-species

The following minor field

Ag[+]

96

A.2. The CHESS database format

molew. = 107.868radius = 1.25e-09element = Agalias = silver

provides for the basis species Ag+, respectively, its (atomic) molecular weight, a radius size (notused in CHESS alone), and its chemical symbol and alias (name).

Redox-couples


HS[-] composition = -2 O2(aq), 1 H[+], 1 SO4[2-]logK = -152.099(0), -138.317(25), -122.137(60), -107.03(100)\

-91.7855(150), -79.4025(200), -69.0257(250), -60.0062(300)radius = 1.75e-10 m

provides, for the redox species HS−, the formation reaction in terms of the basis species, a listof related logK values at different temperatures inC. Note that a radius may be given for eachion. The ion radius is used in CHESS for the B-Dot formulation.

Aqueous-species


MgHCO3[+] composition = 1 HCO3[-], 1 Mg[2+]logK = 1.0798(0), 1.0357(25), 1.1638(60), 1.4355(100)\

1.8804(150), 2.4146(200), 3.0493(250), 3.8424(300)

provides, for the aqueous speciesMgHCO3[+], the formation reaction in terms of the basis species,a list of related logK values at different temperatures inC. An aqueous species is normallyexpressed in terms of the default basis species, with however the major exception of the speciesrelated to redox couples. Using the example ofFeCl[2+], we have

FeCl[2+] composition = 1 Cl[-], 1 Fe[3+]...

97


hence the composition does not refer toFe[2+], but rather toFe[3+] which shares the sameoxidation state asFeCl[2+]. This way of writing the composition allows CHESS to decoupleredox species, if appropriate.

Gaseous-species


CO2(g) composition = -1 H2O, 1 H[+], 1 HCO3[-]logK = 7.6765(0), 7.8136(25), 8.0527(60), 8.3574(100)\

8.7692(150), 9.2165(200), 9.7202(250), 10.3393(300)

provides for the gaseous speciesCO2(g) the formation reaction in terms of the basis species anda list of related logK values for different temperatures inC.

Minerals

Goethite composition = -3 H[+], 1 Fe[3+], 2 H2OlogK = -1.5252(0), -0.5345(25), 0.6(60), 1.6142(100)\

2.6004(150), 3.399(200),vol.weight = 4267.71 kg/m3

provides for the mineral speciesGoethite the formation reaction in terms of the basis (or redox)species, the related log(K) values at different temperatures inC, and the volumic weight. Oneor several sorption sites can be directly defined in the mineral minor field. For example, in thecase of a clay mineral with ion exchange sites, the field

Illite composition = -8 H[+], 0.25 Mg[2+], 0.6 K[+], 2.3 Al[3+], 3.5SiO2(aq), 5 H2OlogK = -11.3859(0), -9.026(25), -5.555(60), -2.0472(100)\1.6128(150), 4.6923(200), 7.4468(250), 10.0976(300)vol.weight = 3000 kg/m3site Illite(Na)

content = 1 Na[+]exch.cap. = 4.5 umol/m2

introduces the siteIllite(Na) which becomes a new basis species, attached to the clay. Thekeywordcontent specifies the species which are susceptible to be released from the site,Na[+]

98

A.2. The CHESS database format

in the example. The exchange capacity (or site density) is expressed inµmol/m2. The numberof exchange sites actually available in a system is calculated as a function of the specific (orvolumetric) surface area given in the input script and the amount of mineral in the system. Insome specific cases the composition can be ignored (or set tounknown) and log(K) skipped.

Colloids


>HFO composition = 3 H2O -3 H[+] 1 Fe[3+]logK = -2.1377vol.weight = 3113.9 kg/m3radius = 10 nmsite >HFO(s)-OH

exch.cap. = 0.093 umol/m2site >HFO(w)-OH

exch.cap. = 3.745 umol/m2

provides the formation reaction of colloidal hydrous ferric oxide in terms of the basis species,a list of related log(K) values at different temperatures inC and its volumetric weight. In ourexample, the model will assume that>HFO is a spherical particle with a radius of 50 nm andcalculates the specific surface area (usingAs= 3/(ρa)), whereρ is the volumic weight anda theradius of the particle. If, however, the colloid is not spherical (e.g. a clay mineral), it becomesmore insidious to provide directly the specific surface areainstead of the radius. The species>HFO may equivalently be entered like this :

>HFO ...vol.weight = 4267.71 kg/m3surface = 14.06 m2/g...

One or several sorption sites can be directly defined in the colloid minor field. In this example,two types of sorption sites (functional groups) are introduced in the second part of the field : astrong site denoted byHFO(s)-OH and a weak site represented by>HFO(w)-OH. Each site is anew basis speciesattachedto its mother-species. The concentration of sites, actually availablein a system is calculated as a function of the specific surfacearea and the total concentration ofcolloids. Note that for a site such asHFO(w)-OH, the keywordcontent may be omitted sinceH[+], the only species which can be released, is always present insolution.

99


An interesting feature of CHESS is the ability to introduce not one colloid size, but merely a sizedistribution. Instead of specifying the radius only, it is possible to define a series of value-pairs :

• the frequency (obtained by SEM counting, for example)

• the radius

Here is an example of colloidal amorphous silica with a size distribution :

>Silica vol.weight = 4826.55 kg/m3radius 0.043 x 30.25 nm0.157 x 33.75 nm0.228 x 37.60 nm0.166 x 41.85 nm0.084 x 46.65 nm0.080 x 51.95 nm0.089 x 57.90 nm0.063 x 64.50 nm0.037 x 71.85 nm0.028 x 80.05 nm0.019 x 89.20 nm0.006 x 99.35 nm

site >Silica-OH exch.cap. = 12 umol/m2

where the first number denotes the frequence (obtained by, e.g., SEM counting), and the secondnumber is the radius corresponding to that frequence. Note that the composition is omitted : asize distribution is not possible in combination with a known composition (i.e., the colloids maynot dissolve).

organic-matter

Theorganic-matter minor fields are, for the moment, identical to the inorganic colloids. Forexample :

>AHA composition = unknownvol.wt. = 1200 kg/m3radius = 5 nmsite >AHA(1)-COOH

exch.cap. = 6 umol/m2molew. = 45.01

site >AHA(2)-COOH

100

A.3. Comments and references

exch.cap. = 3 umol/m2molew. = 45.01

site >AHA-OH

exch.cap. = 1.6 umol/m2molew. = 17.0073

which defines Aldrich humic acid with three different types of functional groups. Note that theelemental composition of a humic acid is rarely known in general. A organics (and solid spe-cies in general) may contain as many sites as needed : it offers henceforth a (not very elegant)possibility to ‘emulate’ the continuous site-distribution model.

Surface-sites


>HFO(s)-OZn[+] composition = 1 >HFO(s)-OH -1 H[+] 1 Zn[2+]logK = 0.99

provides for the site species>HFO(s)-OZn[+] the formation reaction in terms of the basis species(which may be a site) and in this example, a single log(K) value valid for all temperatures.

A.3 Comments and references

The database format allows optionally to provide comments and references for each speciesspecifically. Comments should be directly related to the reaction involved to form the, such asimportant information about the experimental conditions and hypothesis which allowed to deter-mine a formation constant, for instance. Comments and references are shown by JCHESS in thedatabase Analyzer. Here is an example of a species with a comment and a reference :

(UO2)2P2O7(c) composition = 2 UO2[2+], 2 HPO4[2-], -1 H2OlogK = 10.3884(0), 12.2736(25), 13.867(50)\

15.2317(75), 16.4134(100)vol.weight = 3000 kg/m3comment

log K(25 C) uncertainty +/- 1.46, at 95 % CLVan’t Hoff equation used to estimate temperature dependency

reference = NEA TDB, UDATA.tdb downloaded 10/02/2002

101


This example shows a comment for species (UO2)2P2O7(c) provided in a specific field. It is alsopossible to provide the command on a single line, or on multiple lines ended by the line-breakcharacter, “\”.

102

Bibliographie

Allison, J., D. Brown, and K. Novo-Gradac (1991).MINTEQA2/PRODEF2 : A GeochemicalAssessment Model for Environmental Systems. Version 3.0 User’s Manual. EPA/600/3-91/021Athens.

Appelo, C. and D. Postma (1993).Geochemistry, groundwater and pollution. Rotterdam,Holland : A.A. Balkema.

Dzombak, D. and F. Morel (1990).Surface Complexation Modeling. Hydrous Ferric Oxide.New York : John Wiley & Sons.

Lucille, P.-L., A. Burnol, and P. Ollar (2000). Chemtrap : a hydrogeochemicalmodel for reactivetransport in porous media.Hydrol. Proc. 14, 2261–2277.

van der Lee, J. (1997).Modelisation du comportement geochimique et du transport des radio-nucleides en presence de colloıdes. Ph. D. thesis,Ecole des Mines de Paris, Paris, France.

van der Lee, J. (1998, September). Thermodynamic and mathematical concepts of CHESS.Technical Report LHM/RD/98/39, CIG,Ecole des Mines de Paris, Fontainebleau, France.

van der Lee, J., L. De Windt, V. Lagneau, and P. Goblet (2002).Module-oriented modeling ofreactive transport with HYTEC.Computers & Geosciences, in print.

Wolery, T. (1992).EQ3/6 : A software package for geochemical modelling of aqueous systems :package overview and installation guide (version 7.0)(UCRL-MA-110662 PT I ed.). LawrenceLivermore National Laboratory.

103

Bibliographie

104

Index

activitycoefficients, 11correction, 8

activity correction, 51analyzer, voir dtabase25

balance, 7

cation exchange, 67chess

run, 5closed system, voir lberate activities17concentration

set value of, 8, 12units, 12

concentration quantity, 77console, 5

verbose, 5constant capacitance, voir sorptioncontexts, voir editors

database, 83–92analyzer, 25, 66available databases, 83comments, references, 91modify, 25new species, 27using another, 83

density, 7diagram mode, 3, 5, 37

set as default, 5, 37diagrams, 37–43

activity, 41mosaic, 39

purification, 41

pourbaix, 38solubility, 41water stability, 37

dissolutionof calcite, voir kineticsof quartz, 46options, 13, 53

double layer model, voir sorption

editors, 8–12concentration editor, 8, 46contexts, 8, 18, 21items, 10, 18, 22output editor, 21selection, 10, 18, 22titration editor, 18

Ehset free, 17, 19set value of, 10

electroneutrality, 57error messages, 5exclude

colloids, 49, 66groups, 27minerals, 66species, 27

fix activities, 17flush, 66

definition of, 19examples, 19, 67volume, 67

freepH, 61

include

105

INDEX

species, 27, 66items, voir editors

JPlot, 32loading datafiles, 33selection, 33

jplot, 21

Kd approach, 24kinetics, 14, 73–81

active surface, 79calcite dissolution, 74, 79equation, 73law, 14nucleus, 15, 77parameters, 14, 74plagioclase albitization, 75rate constant, 15saturation index, 16temperature effect, 15, 79

liberate activities, 17

mix, 71definition of, 19examples, 31, 71solution, 19, 31, 71with mobile fraction, 71

mobile fraction, 19

nucleus, voir kinetics

open system, voir fx activities17output

examples, 45, 46, 50general report, 28options, voir report optionsreaction results, 28

peset value of, 10

pHfree, 71free early, 59set free, 17, 19

set value of, 10pick up, voir resumepiper diagrams, 30

data format, 31precipitation

of colloids, 54of hematite, 49options, 13, 52, 54

quantityactivity, 11aqueous-concentration, 11concentration, 11definition of, 4, 10fugacity, 11mineral, 11total-concentration, 10

reaction paths, 16–22, 57–72definition of, 16flush, voir flushmix, voir mixtitrate, voir titration

redoxoptions, 8, 54

report, 5report options, 28resume, 28, 72

sampledisplay in report, 28

samples, 16, 19, 31saturation-index, 22script file

load, print, save, 4selection, voir editorssites

define new, 68solvent

activity, 8set mass, 12

sorption, 22speciation mode, 2, 4, 7species

106

INDEX

add, 7define new, 14, 68delete, 7edit, 7toggle on/off, 7

specific surface, 12, 16surface complexation, 23

amended approach, 23electrostatic effects, 23hydrous ferric oxide, 63

tds, 31temperature, 7

reaction with, 66time

initial equilibration, 7titration, 17, 57–66

logarithmic steps, 63with HCl, 61with time, 77with CO2(g), 58with H2O, 18with pH, 64

total-concentration, voir quantitytriple layer model, voir sorption

volume, 7, 19, 67

107

INDEX

108

About JCHESS

CHESS is a complete geochemical model to simulate the equilibrium state of complex aquaticsystems, including minerals, organics, colloids and gases. The availability of JCHESS, a graphicuser’s interface to CHESS, greatly facilitates the use and brings even complex options to thereach of all modelling scientists and engineers. The current version includes (among others) thefollowing processes and features :

• thermodynamic equilibrium of mineral, colloidal, organic, gaseous or aqueous species

• calculation of Pourbaix, activity and solubility diagrams

• oxidation and reduction processes

• precipitation and dissolution of mineral phases

• formation and dissolution of colloidal phases (organic, inorganic)

• all common surface interface reactions

• electrostatic correction models

• multi-site and multi-surface reactions

• several reaction path models (titration, mixing, flushing)

• temperature dependence

• complete kinetic control of dissolution and precipitation

• heterogeneous size distributions of colloids

CHESS is designed to meet demands of the most complex geochemical modelling needs, yetit also aims to meet requirements of modern user-friendliness. About seven different thermody-namic databases have been made available for use with CHESS and they are readily extendedwith new species. The numerous options of the code are clearly exposed and powerful editorsare provided to manipulate the input requirements in a fail-save manner. The graphical front-end of JCHESS is a valuable help for datafitting and allows to construct complex Pourbaix-typediagrams.

The CHESS Web site :http://chess.ensmp.frE-Mail : [email protected]

All software and design created by J. van der Leec©1993 – 2002 ARMINES

Documents

Jchess Tutorial 3.0